{
  "best_metric": 0.31124767661094666,
  "best_model_checkpoint": "experts/mistralic-expert-17/checkpoint-4800",
  "epoch": 1.1606819006166122,
  "eval_steps": 200,
  "global_step": 5600,
  "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": 0.3764,
      "step": 1
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.3983,
      "step": 2
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.4823,
      "step": 3
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.4139,
      "step": 4
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.341,
      "step": 5
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.3615,
      "step": 6
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.3709,
      "step": 7
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.3184,
      "step": 8
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.3923,
      "step": 9
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.5588,
      "step": 10
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.4881,
      "step": 11
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.4644,
      "step": 12
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.3299,
      "step": 13
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.404,
      "step": 14
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.3933,
      "step": 15
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.3784,
      "step": 16
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.406,
      "step": 17
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.4184,
      "step": 18
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.3953,
      "step": 19
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.4085,
      "step": 20
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.4547,
      "step": 21
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.4166,
      "step": 22
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.4566,
      "step": 23
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.3928,
      "step": 24
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3639,
      "step": 25
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.386,
      "step": 26
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3157,
      "step": 27
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3111,
      "step": 28
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5034,
      "step": 29
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4071,
      "step": 30
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6236,
      "step": 31
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3294,
      "step": 32
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4722,
      "step": 33
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3625,
      "step": 34
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5145,
      "step": 35
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.416,
      "step": 36
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3519,
      "step": 37
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3509,
      "step": 38
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.416,
      "step": 39
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4899,
      "step": 40
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3193,
      "step": 41
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5121,
      "step": 42
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3911,
      "step": 43
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4805,
      "step": 44
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3315,
      "step": 45
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3784,
      "step": 46
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4312,
      "step": 47
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4413,
      "step": 48
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3399,
      "step": 49
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3861,
      "step": 50
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4977,
      "step": 51
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.515,
      "step": 52
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3766,
      "step": 53
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3144,
      "step": 54
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3097,
      "step": 55
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3277,
      "step": 56
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3147,
      "step": 57
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4264,
      "step": 58
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3957,
      "step": 59
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3379,
      "step": 60
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3163,
      "step": 61
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3506,
      "step": 62
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3888,
      "step": 63
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.347,
      "step": 64
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.522,
      "step": 65
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3002,
      "step": 66
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3523,
      "step": 67
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6818,
      "step": 68
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4075,
      "step": 69
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4022,
      "step": 70
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4122,
      "step": 71
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4788,
      "step": 72
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3144,
      "step": 73
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.4427,
      "step": 74
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.453,
      "step": 75
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3033,
      "step": 76
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.475,
      "step": 77
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.4291,
      "step": 78
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.4268,
      "step": 79
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3052,
      "step": 80
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.2165,
      "step": 81
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5203,
      "step": 82
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3173,
      "step": 83
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3227,
      "step": 84
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.387,
      "step": 85
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3191,
      "step": 86
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.2438,
      "step": 87
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3296,
      "step": 88
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.2889,
      "step": 89
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3352,
      "step": 90
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3428,
      "step": 91
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3576,
      "step": 92
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3154,
      "step": 93
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.4834,
      "step": 94
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.2879,
      "step": 95
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3734,
      "step": 96
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.4016,
      "step": 97
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.4421,
      "step": 98
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3338,
      "step": 99
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.4342,
      "step": 100
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3754,
      "step": 101
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5515,
      "step": 102
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3301,
      "step": 103
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3824,
      "step": 104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3394,
      "step": 105
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.2945,
      "step": 106
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3857,
      "step": 107
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3952,
      "step": 108
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3497,
      "step": 109
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.4295,
      "step": 110
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3641,
      "step": 111
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.2941,
      "step": 112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.4429,
      "step": 113
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.2982,
      "step": 114
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.407,
      "step": 115
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5319,
      "step": 116
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3071,
      "step": 117
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.353,
      "step": 118
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5562,
      "step": 119
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3285,
      "step": 120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3044,
      "step": 121
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3912,
      "step": 122
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3579,
      "step": 123
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4041,
      "step": 124
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4563,
      "step": 125
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3469,
      "step": 126
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3381,
      "step": 127
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3378,
      "step": 128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4174,
      "step": 129
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3629,
      "step": 130
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3426,
      "step": 131
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5591,
      "step": 132
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.321,
      "step": 133
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3866,
      "step": 134
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4362,
      "step": 135
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3987,
      "step": 136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4104,
      "step": 137
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3393,
      "step": 138
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3664,
      "step": 139
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.2806,
      "step": 140
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5358,
      "step": 141
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5086,
      "step": 142
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3179,
      "step": 143
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.2922,
      "step": 144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.413,
      "step": 145
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3515,
      "step": 146
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3137,
      "step": 147
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3654,
      "step": 148
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5826,
      "step": 149
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4276,
      "step": 150
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.2985,
      "step": 151
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3914,
      "step": 152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.2938,
      "step": 153
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.2313,
      "step": 154
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4092,
      "step": 155
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4689,
      "step": 156
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3192,
      "step": 157
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3818,
      "step": 158
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3851,
      "step": 159
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.2992,
      "step": 160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3687,
      "step": 161
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3448,
      "step": 162
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3684,
      "step": 163
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.2847,
      "step": 164
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3367,
      "step": 165
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.2416,
      "step": 166
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4329,
      "step": 167
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4203,
      "step": 168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.355,
      "step": 169
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3465,
      "step": 170
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.2676,
      "step": 171
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3554,
      "step": 172
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.2724,
      "step": 173
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4885,
      "step": 174
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.556,
      "step": 175
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.298,
      "step": 176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.286,
      "step": 177
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3485,
      "step": 178
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3193,
      "step": 179
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3288,
      "step": 180
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.2939,
      "step": 181
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3689,
      "step": 182
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.324,
      "step": 183
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3524,
      "step": 184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3494,
      "step": 185
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3723,
      "step": 186
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3995,
      "step": 187
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4912,
      "step": 188
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3704,
      "step": 189
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3676,
      "step": 190
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4459,
      "step": 191
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.336,
      "step": 192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.627,
      "step": 193
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5047,
      "step": 194
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3175,
      "step": 195
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4015,
      "step": 196
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3854,
      "step": 197
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3298,
      "step": 198
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3249,
      "step": 199
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3183,
      "step": 200
    },
    {
      "epoch": 0.04,
      "eval_loss": 0.36924177408218384,
      "eval_runtime": 159.4711,
      "eval_samples_per_second": 6.271,
      "eval_steps_per_second": 3.135,
      "step": 200
    },
    {
      "epoch": 0.04,
      "mmlu_eval_accuracy": 0.5940031246983312,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.35714285714285715,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "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": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.5454545454545454,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.6666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.43902439024390244,
      "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": 0.7222222222222222,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5348837209302325,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3103448275862069,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6923076923076923,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
      "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": 0.7391304347826086,
      "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": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.37,
      "mmlu_eval_accuracy_nutrition": 0.7878787878787878,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.4857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_law": 0.4,
      "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": 0.7037037037037037,
      "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": 0.8947368421052632,
      "mmlu_loss": 1.4302824790888604,
      "step": 200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4524,
      "step": 201
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4312,
      "step": 202
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.2717,
      "step": 203
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3977,
      "step": 204
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3864,
      "step": 205
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.2953,
      "step": 206
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3302,
      "step": 207
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4244,
      "step": 208
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4016,
      "step": 209
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4256,
      "step": 210
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4162,
      "step": 211
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.282,
      "step": 212
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3207,
      "step": 213
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.2938,
      "step": 214
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.2823,
      "step": 215
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.2986,
      "step": 216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.2869,
      "step": 217
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.317,
      "step": 218
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2975,
      "step": 219
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2523,
      "step": 220
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3691,
      "step": 221
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.341,
      "step": 222
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2603,
      "step": 223
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.4282,
      "step": 224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.4877,
      "step": 225
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2787,
      "step": 226
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3901,
      "step": 227
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3039,
      "step": 228
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.4518,
      "step": 229
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.325,
      "step": 230
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2634,
      "step": 231
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3732,
      "step": 232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3162,
      "step": 233
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3761,
      "step": 234
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2739,
      "step": 235
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.344,
      "step": 236
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2995,
      "step": 237
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2882,
      "step": 238
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2372,
      "step": 239
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3331,
      "step": 240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.6815,
      "step": 241
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2672,
      "step": 242
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.4416,
      "step": 243
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2813,
      "step": 244
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3574,
      "step": 245
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3237,
      "step": 246
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2798,
      "step": 247
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5538,
      "step": 248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3436,
      "step": 249
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3082,
      "step": 250
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5797,
      "step": 251
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2435,
      "step": 252
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.4687,
      "step": 253
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2921,
      "step": 254
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3151,
      "step": 255
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2787,
      "step": 256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3544,
      "step": 257
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2891,
      "step": 258
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.4261,
      "step": 259
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2829,
      "step": 260
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.2342,
      "step": 261
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3482,
      "step": 262
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3608,
      "step": 263
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.4406,
      "step": 264
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3543,
      "step": 265
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4826,
      "step": 266
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3967,
      "step": 267
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3463,
      "step": 268
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.2837,
      "step": 269
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3827,
      "step": 270
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3583,
      "step": 271
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3035,
      "step": 272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3266,
      "step": 273
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5502,
      "step": 274
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4045,
      "step": 275
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.2922,
      "step": 276
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4068,
      "step": 277
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.32,
      "step": 278
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3039,
      "step": 279
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5073,
      "step": 280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.357,
      "step": 281
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.2757,
      "step": 282
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3473,
      "step": 283
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4409,
      "step": 284
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4095,
      "step": 285
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3177,
      "step": 286
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5387,
      "step": 287
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3647,
      "step": 288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4568,
      "step": 289
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4054,
      "step": 290
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3302,
      "step": 291
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3426,
      "step": 292
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4874,
      "step": 293
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5772,
      "step": 294
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.2785,
      "step": 295
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3493,
      "step": 296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3909,
      "step": 297
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4654,
      "step": 298
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4774,
      "step": 299
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.2985,
      "step": 300
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4445,
      "step": 301
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3078,
      "step": 302
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.2923,
      "step": 303
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.373,
      "step": 304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3412,
      "step": 305
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.2629,
      "step": 306
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.258,
      "step": 307
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4849,
      "step": 308
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4029,
      "step": 309
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3758,
      "step": 310
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4228,
      "step": 311
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3841,
      "step": 312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3038,
      "step": 313
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4246,
      "step": 314
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3501,
      "step": 315
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3886,
      "step": 316
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3705,
      "step": 317
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3549,
      "step": 318
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3336,
      "step": 319
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3522,
      "step": 320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.2667,
      "step": 321
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4979,
      "step": 322
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3397,
      "step": 323
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3273,
      "step": 324
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3509,
      "step": 325
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3015,
      "step": 326
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.2906,
      "step": 327
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.2315,
      "step": 328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3104,
      "step": 329
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4869,
      "step": 330
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4246,
      "step": 331
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3485,
      "step": 332
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3974,
      "step": 333
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4052,
      "step": 334
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4124,
      "step": 335
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4349,
      "step": 336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3406,
      "step": 337
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3532,
      "step": 338
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4889,
      "step": 339
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5194,
      "step": 340
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.514,
      "step": 341
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.327,
      "step": 342
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3088,
      "step": 343
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3761,
      "step": 344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.2712,
      "step": 345
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3488,
      "step": 346
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4925,
      "step": 347
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3292,
      "step": 348
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.282,
      "step": 349
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3427,
      "step": 350
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3182,
      "step": 351
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4053,
      "step": 352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3911,
      "step": 353
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.2906,
      "step": 354
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4012,
      "step": 355
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3397,
      "step": 356
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3832,
      "step": 357
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3739,
      "step": 358
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3106,
      "step": 359
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3123,
      "step": 360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3197,
      "step": 361
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3987,
      "step": 362
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.2746,
      "step": 363
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3073,
      "step": 364
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3593,
      "step": 365
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3287,
      "step": 366
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.354,
      "step": 367
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3877,
      "step": 368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.4417,
      "step": 369
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.2696,
      "step": 370
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3225,
      "step": 371
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3691,
      "step": 372
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3359,
      "step": 373
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.2475,
      "step": 374
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3173,
      "step": 375
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3585,
      "step": 376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.2903,
      "step": 377
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3593,
      "step": 378
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3115,
      "step": 379
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5,
      "step": 380
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.4372,
      "step": 381
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.319,
      "step": 382
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.2573,
      "step": 383
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5916,
      "step": 384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.4126,
      "step": 385
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3318,
      "step": 386
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.2875,
      "step": 387
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.2891,
      "step": 388
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3617,
      "step": 389
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3148,
      "step": 390
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.4457,
      "step": 391
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.407,
      "step": 392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.4488,
      "step": 393
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.4066,
      "step": 394
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3808,
      "step": 395
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.2792,
      "step": 396
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.4491,
      "step": 397
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3646,
      "step": 398
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3868,
      "step": 399
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.2754,
      "step": 400
    },
    {
      "epoch": 0.08,
      "eval_loss": 0.358918160200119,
      "eval_runtime": 159.4642,
      "eval_samples_per_second": 6.271,
      "eval_steps_per_second": 3.136,
      "step": 400
    },
    {
      "epoch": 0.08,
      "mmlu_eval_accuracy": 0.5999132037466034,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "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": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.6666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.5625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4878048780487805,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.59375,
      "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": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3448275862068966,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6923076923076923,
      "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": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
      "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": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7209302325581395,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.33,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.42857142857142855,
      "mmlu_eval_accuracy_professional_accounting": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_law": 0.3941176470588235,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.6376811594202898,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "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": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.4201551996774524,
      "step": 400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3131,
      "step": 401
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3419,
      "step": 402
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3984,
      "step": 403
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3459,
      "step": 404
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.4061,
      "step": 405
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.2641,
      "step": 406
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.332,
      "step": 407
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3048,
      "step": 408
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3402,
      "step": 409
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3617,
      "step": 410
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4099,
      "step": 411
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6116,
      "step": 412
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4472,
      "step": 413
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.2897,
      "step": 414
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3139,
      "step": 415
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4072,
      "step": 416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3577,
      "step": 417
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4119,
      "step": 418
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3019,
      "step": 419
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5519,
      "step": 420
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.2923,
      "step": 421
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4213,
      "step": 422
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3394,
      "step": 423
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3897,
      "step": 424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.2891,
      "step": 425
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.285,
      "step": 426
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3758,
      "step": 427
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.2731,
      "step": 428
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.374,
      "step": 429
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3646,
      "step": 430
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4202,
      "step": 431
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.2353,
      "step": 432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.383,
      "step": 433
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3019,
      "step": 434
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5666,
      "step": 435
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3783,
      "step": 436
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.2525,
      "step": 437
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3956,
      "step": 438
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.395,
      "step": 439
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.2849,
      "step": 440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3503,
      "step": 441
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3204,
      "step": 442
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3204,
      "step": 443
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3083,
      "step": 444
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4108,
      "step": 445
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.2786,
      "step": 446
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5366,
      "step": 447
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.2785,
      "step": 448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3132,
      "step": 449
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4062,
      "step": 450
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3812,
      "step": 451
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.2555,
      "step": 452
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4278,
      "step": 453
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.2947,
      "step": 454
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3247,
      "step": 455
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5202,
      "step": 456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3669,
      "step": 457
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3395,
      "step": 458
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.358,
      "step": 459
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.2811,
      "step": 460
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4165,
      "step": 461
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.2591,
      "step": 462
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3491,
      "step": 463
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.33,
      "step": 464
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4262,
      "step": 465
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3111,
      "step": 466
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4079,
      "step": 467
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4417,
      "step": 468
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.261,
      "step": 469
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.2752,
      "step": 470
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.2719,
      "step": 471
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3703,
      "step": 472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.2798,
      "step": 473
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3553,
      "step": 474
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3205,
      "step": 475
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3406,
      "step": 476
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3759,
      "step": 477
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3276,
      "step": 478
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3335,
      "step": 479
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5614,
      "step": 480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.2922,
      "step": 481
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3785,
      "step": 482
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3085,
      "step": 483
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4565,
      "step": 484
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4444,
      "step": 485
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.2936,
      "step": 486
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3858,
      "step": 487
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3834,
      "step": 488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3273,
      "step": 489
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3334,
      "step": 490
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3236,
      "step": 491
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.2944,
      "step": 492
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5452,
      "step": 493
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3507,
      "step": 494
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3539,
      "step": 495
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.2818,
      "step": 496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4009,
      "step": 497
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3196,
      "step": 498
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4456,
      "step": 499
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.2854,
      "step": 500
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.359,
      "step": 501
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3794,
      "step": 502
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.581,
      "step": 503
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3004,
      "step": 504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3796,
      "step": 505
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3091,
      "step": 506
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4228,
      "step": 507
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3734,
      "step": 508
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.2583,
      "step": 509
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3844,
      "step": 510
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3349,
      "step": 511
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.347,
      "step": 512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.2649,
      "step": 513
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3511,
      "step": 514
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.1902,
      "step": 515
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.2628,
      "step": 516
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.258,
      "step": 517
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4467,
      "step": 518
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3565,
      "step": 519
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3713,
      "step": 520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.2867,
      "step": 521
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.2861,
      "step": 522
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.2694,
      "step": 523
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4202,
      "step": 524
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4187,
      "step": 525
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4243,
      "step": 526
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4473,
      "step": 527
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4199,
      "step": 528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3368,
      "step": 529
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3606,
      "step": 530
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6544,
      "step": 531
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3448,
      "step": 532
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3672,
      "step": 533
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3245,
      "step": 534
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3069,
      "step": 535
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.47,
      "step": 536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3799,
      "step": 537
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3455,
      "step": 538
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4227,
      "step": 539
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3065,
      "step": 540
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3319,
      "step": 541
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5573,
      "step": 542
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3221,
      "step": 543
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.2999,
      "step": 544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4778,
      "step": 545
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4276,
      "step": 546
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3707,
      "step": 547
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3303,
      "step": 548
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5229,
      "step": 549
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3393,
      "step": 550
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3118,
      "step": 551
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.2864,
      "step": 552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3364,
      "step": 553
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3086,
      "step": 554
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3338,
      "step": 555
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.2945,
      "step": 556
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3808,
      "step": 557
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.2725,
      "step": 558
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.2995,
      "step": 559
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3137,
      "step": 560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3209,
      "step": 561
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.2393,
      "step": 562
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3013,
      "step": 563
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.2315,
      "step": 564
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3284,
      "step": 565
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3422,
      "step": 566
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4019,
      "step": 567
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5752,
      "step": 568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.317,
      "step": 569
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3557,
      "step": 570
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3452,
      "step": 571
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3375,
      "step": 572
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.2105,
      "step": 573
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4988,
      "step": 574
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.2949,
      "step": 575
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3991,
      "step": 576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3326,
      "step": 577
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3207,
      "step": 578
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3678,
      "step": 579
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.2938,
      "step": 580
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3795,
      "step": 581
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4644,
      "step": 582
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4052,
      "step": 583
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.2996,
      "step": 584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.455,
      "step": 585
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4427,
      "step": 586
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3916,
      "step": 587
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3169,
      "step": 588
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.576,
      "step": 589
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4151,
      "step": 590
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.2811,
      "step": 591
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3179,
      "step": 592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.2883,
      "step": 593
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5965,
      "step": 594
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3987,
      "step": 595
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4434,
      "step": 596
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5114,
      "step": 597
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.2867,
      "step": 598
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.255,
      "step": 599
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.33,
      "step": 600
    },
    {
      "epoch": 0.12,
      "eval_loss": 0.3515583574771881,
      "eval_runtime": 159.3504,
      "eval_samples_per_second": 6.275,
      "eval_steps_per_second": 3.138,
      "step": 600
    },
    {
      "epoch": 0.12,
      "mmlu_eval_accuracy": 0.6095696896582704,
      "mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.8125,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.5454545454545454,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.6666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.5625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4634146341463415,
      "mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3448275862068966,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.85,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
      "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": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7558139534883721,
      "mmlu_eval_accuracy_moral_disputes": 0.6052631578947368,
      "mmlu_eval_accuracy_moral_scenarios": 0.36,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.6764705882352942,
      "mmlu_eval_accuracy_prehistory": 0.5428571428571428,
      "mmlu_eval_accuracy_professional_accounting": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_law": 0.4,
      "mmlu_eval_accuracy_professional_medicine": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6666666666666666,
      "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": 0.7894736842105263,
      "mmlu_loss": 1.383797744780234,
      "step": 600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3336,
      "step": 601
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3879,
      "step": 602
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3637,
      "step": 603
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3342,
      "step": 604
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.2545,
      "step": 605
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.478,
      "step": 606
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3672,
      "step": 607
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3338,
      "step": 608
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.2637,
      "step": 609
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5447,
      "step": 610
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.2509,
      "step": 611
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3597,
      "step": 612
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3357,
      "step": 613
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.2906,
      "step": 614
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3108,
      "step": 615
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4441,
      "step": 616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4973,
      "step": 617
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4386,
      "step": 618
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.2435,
      "step": 619
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3177,
      "step": 620
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.2406,
      "step": 621
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.344,
      "step": 622
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.519,
      "step": 623
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4641,
      "step": 624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3593,
      "step": 625
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.442,
      "step": 626
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.2441,
      "step": 627
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3301,
      "step": 628
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3552,
      "step": 629
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3342,
      "step": 630
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3023,
      "step": 631
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4696,
      "step": 632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3121,
      "step": 633
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3117,
      "step": 634
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3381,
      "step": 635
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4275,
      "step": 636
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5997,
      "step": 637
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3677,
      "step": 638
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3019,
      "step": 639
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3838,
      "step": 640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5191,
      "step": 641
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.2501,
      "step": 642
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.2363,
      "step": 643
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3212,
      "step": 644
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4166,
      "step": 645
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3708,
      "step": 646
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3446,
      "step": 647
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.2667,
      "step": 648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3117,
      "step": 649
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4283,
      "step": 650
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3815,
      "step": 651
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3552,
      "step": 652
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2827,
      "step": 653
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5081,
      "step": 654
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3505,
      "step": 655
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3409,
      "step": 656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4335,
      "step": 657
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3517,
      "step": 658
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4015,
      "step": 659
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2506,
      "step": 660
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2683,
      "step": 661
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3261,
      "step": 662
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3973,
      "step": 663
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3333,
      "step": 664
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2956,
      "step": 665
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4357,
      "step": 666
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3456,
      "step": 667
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2607,
      "step": 668
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2917,
      "step": 669
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5605,
      "step": 670
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2892,
      "step": 671
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4714,
      "step": 672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3037,
      "step": 673
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4018,
      "step": 674
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.29,
      "step": 675
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3111,
      "step": 676
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3386,
      "step": 677
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2688,
      "step": 678
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2314,
      "step": 679
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2978,
      "step": 680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4678,
      "step": 681
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2993,
      "step": 682
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3585,
      "step": 683
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2585,
      "step": 684
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3749,
      "step": 685
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5372,
      "step": 686
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6269,
      "step": 687
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3462,
      "step": 688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3386,
      "step": 689
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3011,
      "step": 690
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3638,
      "step": 691
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4762,
      "step": 692
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.2494,
      "step": 693
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.25,
      "step": 694
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3209,
      "step": 695
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3363,
      "step": 696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3057,
      "step": 697
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.29,
      "step": 698
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.279,
      "step": 699
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4144,
      "step": 700
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4035,
      "step": 701
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4001,
      "step": 702
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3917,
      "step": 703
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4514,
      "step": 704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.275,
      "step": 705
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4256,
      "step": 706
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3787,
      "step": 707
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4171,
      "step": 708
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3542,
      "step": 709
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.2872,
      "step": 710
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3524,
      "step": 711
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4246,
      "step": 712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.2009,
      "step": 713
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3558,
      "step": 714
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4853,
      "step": 715
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3248,
      "step": 716
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3283,
      "step": 717
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3036,
      "step": 718
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.2898,
      "step": 719
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.2929,
      "step": 720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3195,
      "step": 721
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3062,
      "step": 722
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3071,
      "step": 723
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3382,
      "step": 724
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3026,
      "step": 725
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.2771,
      "step": 726
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3467,
      "step": 727
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3839,
      "step": 728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.2969,
      "step": 729
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.2924,
      "step": 730
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4699,
      "step": 731
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.2693,
      "step": 732
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3428,
      "step": 733
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3306,
      "step": 734
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4782,
      "step": 735
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4112,
      "step": 736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3568,
      "step": 737
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3454,
      "step": 738
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3465,
      "step": 739
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3941,
      "step": 740
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4345,
      "step": 741
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.234,
      "step": 742
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3432,
      "step": 743
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3649,
      "step": 744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4051,
      "step": 745
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3145,
      "step": 746
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3366,
      "step": 747
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2769,
      "step": 748
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.438,
      "step": 749
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.4599,
      "step": 750
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.465,
      "step": 751
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2447,
      "step": 752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3375,
      "step": 753
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.356,
      "step": 754
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3919,
      "step": 755
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.4425,
      "step": 756
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2936,
      "step": 757
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3871,
      "step": 758
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3868,
      "step": 759
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3073,
      "step": 760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3297,
      "step": 761
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3452,
      "step": 762
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2743,
      "step": 763
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.4263,
      "step": 764
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.263,
      "step": 765
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.249,
      "step": 766
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3263,
      "step": 767
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3864,
      "step": 768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3744,
      "step": 769
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3379,
      "step": 770
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2812,
      "step": 771
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2239,
      "step": 772
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2832,
      "step": 773
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3457,
      "step": 774
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3176,
      "step": 775
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3036,
      "step": 776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3842,
      "step": 777
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2456,
      "step": 778
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.373,
      "step": 779
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.411,
      "step": 780
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2447,
      "step": 781
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.262,
      "step": 782
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.331,
      "step": 783
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2911,
      "step": 784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2906,
      "step": 785
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3722,
      "step": 786
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2862,
      "step": 787
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.4915,
      "step": 788
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3048,
      "step": 789
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3548,
      "step": 790
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2886,
      "step": 791
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2657,
      "step": 792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.4349,
      "step": 793
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3265,
      "step": 794
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2918,
      "step": 795
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.2429,
      "step": 796
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3048,
      "step": 797
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3035,
      "step": 798
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3287,
      "step": 799
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2949,
      "step": 800
    },
    {
      "epoch": 0.17,
      "eval_loss": 0.3455464243888855,
      "eval_runtime": 159.257,
      "eval_samples_per_second": 6.279,
      "eval_steps_per_second": 3.14,
      "step": 800
    },
    {
      "epoch": 0.17,
      "mmlu_eval_accuracy": 0.5986589042634419,
      "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": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.75,
      "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": 0.5625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
      "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": 0.3103448275862069,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "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": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.36,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.4857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_law": 0.4,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.6086956521739131,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6666666666666666,
      "mmlu_eval_accuracy_sociology": 0.8636363636363636,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.6111111111111112,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.5759579020287287,
      "step": 800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4054,
      "step": 801
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4718,
      "step": 802
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2695,
      "step": 803
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3716,
      "step": 804
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4109,
      "step": 805
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3176,
      "step": 806
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4368,
      "step": 807
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3341,
      "step": 808
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3081,
      "step": 809
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2677,
      "step": 810
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3201,
      "step": 811
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2704,
      "step": 812
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3967,
      "step": 813
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.8647,
      "step": 814
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3318,
      "step": 815
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2868,
      "step": 816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4796,
      "step": 817
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3111,
      "step": 818
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2656,
      "step": 819
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2656,
      "step": 820
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2475,
      "step": 821
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.388,
      "step": 822
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2936,
      "step": 823
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2784,
      "step": 824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3155,
      "step": 825
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.1861,
      "step": 826
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2374,
      "step": 827
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2553,
      "step": 828
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2571,
      "step": 829
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3295,
      "step": 830
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4078,
      "step": 831
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2489,
      "step": 832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2886,
      "step": 833
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2767,
      "step": 834
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6366,
      "step": 835
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4288,
      "step": 836
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2506,
      "step": 837
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4165,
      "step": 838
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.2654,
      "step": 839
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3846,
      "step": 840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3165,
      "step": 841
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3183,
      "step": 842
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.329,
      "step": 843
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3764,
      "step": 844
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3789,
      "step": 845
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.2707,
      "step": 846
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4665,
      "step": 847
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3743,
      "step": 848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3822,
      "step": 849
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.2279,
      "step": 850
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3113,
      "step": 851
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3191,
      "step": 852
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3549,
      "step": 853
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3279,
      "step": 854
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3831,
      "step": 855
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.2911,
      "step": 856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5018,
      "step": 857
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.2493,
      "step": 858
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3465,
      "step": 859
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4005,
      "step": 860
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3605,
      "step": 861
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4068,
      "step": 862
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3948,
      "step": 863
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3638,
      "step": 864
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3076,
      "step": 865
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3566,
      "step": 866
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4456,
      "step": 867
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3633,
      "step": 868
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5587,
      "step": 869
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5333,
      "step": 870
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3908,
      "step": 871
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.2926,
      "step": 872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3232,
      "step": 873
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4943,
      "step": 874
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3418,
      "step": 875
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3492,
      "step": 876
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3138,
      "step": 877
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3047,
      "step": 878
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.2521,
      "step": 879
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.2705,
      "step": 880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3355,
      "step": 881
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3358,
      "step": 882
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3186,
      "step": 883
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3052,
      "step": 884
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.2069,
      "step": 885
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3336,
      "step": 886
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.2978,
      "step": 887
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.2614,
      "step": 888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.2852,
      "step": 889
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.324,
      "step": 890
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4048,
      "step": 891
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3176,
      "step": 892
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4053,
      "step": 893
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2959,
      "step": 894
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3648,
      "step": 895
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3507,
      "step": 896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3496,
      "step": 897
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2629,
      "step": 898
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3048,
      "step": 899
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3152,
      "step": 900
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3095,
      "step": 901
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3185,
      "step": 902
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3165,
      "step": 903
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3354,
      "step": 904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3616,
      "step": 905
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4569,
      "step": 906
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2608,
      "step": 907
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2492,
      "step": 908
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2194,
      "step": 909
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3576,
      "step": 910
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3949,
      "step": 911
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4273,
      "step": 912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3187,
      "step": 913
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4265,
      "step": 914
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.375,
      "step": 915
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4456,
      "step": 916
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2682,
      "step": 917
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3754,
      "step": 918
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3388,
      "step": 919
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4208,
      "step": 920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2531,
      "step": 921
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3969,
      "step": 922
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2425,
      "step": 923
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2948,
      "step": 924
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3525,
      "step": 925
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2286,
      "step": 926
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2691,
      "step": 927
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2962,
      "step": 928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3169,
      "step": 929
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3709,
      "step": 930
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2728,
      "step": 931
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4006,
      "step": 932
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2902,
      "step": 933
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3399,
      "step": 934
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4129,
      "step": 935
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3946,
      "step": 936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3531,
      "step": 937
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5031,
      "step": 938
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.2395,
      "step": 939
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4327,
      "step": 940
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3431,
      "step": 941
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3588,
      "step": 942
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2331,
      "step": 943
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3002,
      "step": 944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2886,
      "step": 945
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3718,
      "step": 946
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4566,
      "step": 947
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3216,
      "step": 948
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4049,
      "step": 949
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5468,
      "step": 950
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2624,
      "step": 951
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3018,
      "step": 952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.328,
      "step": 953
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3149,
      "step": 954
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.293,
      "step": 955
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3949,
      "step": 956
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2829,
      "step": 957
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5452,
      "step": 958
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3265,
      "step": 959
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3454,
      "step": 960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2472,
      "step": 961
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3075,
      "step": 962
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3953,
      "step": 963
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4237,
      "step": 964
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3237,
      "step": 965
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.224,
      "step": 966
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3196,
      "step": 967
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3051,
      "step": 968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2262,
      "step": 969
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3367,
      "step": 970
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2836,
      "step": 971
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2292,
      "step": 972
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3596,
      "step": 973
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.251,
      "step": 974
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3142,
      "step": 975
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2804,
      "step": 976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2589,
      "step": 977
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2276,
      "step": 978
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.247,
      "step": 979
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3212,
      "step": 980
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2639,
      "step": 981
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4801,
      "step": 982
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3661,
      "step": 983
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2641,
      "step": 984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4465,
      "step": 985
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2491,
      "step": 986
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3758,
      "step": 987
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3068,
      "step": 988
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2582,
      "step": 989
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3422,
      "step": 990
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3677,
      "step": 991
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2968,
      "step": 992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2564,
      "step": 993
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2396,
      "step": 994
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5021,
      "step": 995
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3229,
      "step": 996
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3049,
      "step": 997
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3978,
      "step": 998
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5366,
      "step": 999
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2309,
      "step": 1000
    },
    {
      "epoch": 0.21,
      "eval_loss": 0.34240442514419556,
      "eval_runtime": 159.3616,
      "eval_samples_per_second": 6.275,
      "eval_steps_per_second": 3.138,
      "step": 1000
    },
    {
      "epoch": 0.21,
      "mmlu_eval_accuracy": 0.6036630168245939,
      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "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": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.6363636363636364,
      "mmlu_eval_accuracy_college_medicine": 0.6818181818181818,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.6538461538461539,
      "mmlu_eval_accuracy_econometrics": 0.6666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.43902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "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": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.4482758620689655,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5769230769230769,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "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": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.6666666666666666,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.36,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.6764705882352942,
      "mmlu_eval_accuracy_prehistory": 0.5714285714285714,
      "mmlu_eval_accuracy_professional_accounting": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_law": 0.40588235294117647,
      "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": 0.6666666666666666,
      "mmlu_eval_accuracy_sociology": 0.9545454545454546,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.6111111111111112,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.4610189728964091,
      "step": 1000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4934,
      "step": 1001
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2727,
      "step": 1002
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2757,
      "step": 1003
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4271,
      "step": 1004
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2999,
      "step": 1005
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2883,
      "step": 1006
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3823,
      "step": 1007
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4903,
      "step": 1008
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3183,
      "step": 1009
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.397,
      "step": 1010
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3115,
      "step": 1011
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4355,
      "step": 1012
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.274,
      "step": 1013
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4141,
      "step": 1014
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.418,
      "step": 1015
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2875,
      "step": 1016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2975,
      "step": 1017
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3519,
      "step": 1018
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2645,
      "step": 1019
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3193,
      "step": 1020
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2065,
      "step": 1021
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3574,
      "step": 1022
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3843,
      "step": 1023
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3478,
      "step": 1024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3781,
      "step": 1025
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2993,
      "step": 1026
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2651,
      "step": 1027
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3909,
      "step": 1028
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5299,
      "step": 1029
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4472,
      "step": 1030
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3763,
      "step": 1031
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4319,
      "step": 1032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2601,
      "step": 1033
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2539,
      "step": 1034
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2708,
      "step": 1035
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2716,
      "step": 1036
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.2905,
      "step": 1037
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3437,
      "step": 1038
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3777,
      "step": 1039
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3246,
      "step": 1040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3998,
      "step": 1041
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2726,
      "step": 1042
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3591,
      "step": 1043
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2694,
      "step": 1044
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3669,
      "step": 1045
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.4534,
      "step": 1046
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2806,
      "step": 1047
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.348,
      "step": 1048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3629,
      "step": 1049
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2536,
      "step": 1050
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3034,
      "step": 1051
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2936,
      "step": 1052
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3849,
      "step": 1053
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3761,
      "step": 1054
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.343,
      "step": 1055
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2749,
      "step": 1056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.329,
      "step": 1057
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3792,
      "step": 1058
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3433,
      "step": 1059
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2232,
      "step": 1060
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5412,
      "step": 1061
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.453,
      "step": 1062
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2917,
      "step": 1063
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2818,
      "step": 1064
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2746,
      "step": 1065
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3564,
      "step": 1066
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.298,
      "step": 1067
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2601,
      "step": 1068
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.4126,
      "step": 1069
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3852,
      "step": 1070
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2738,
      "step": 1071
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3528,
      "step": 1072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3714,
      "step": 1073
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3275,
      "step": 1074
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.436,
      "step": 1075
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.2869,
      "step": 1076
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3169,
      "step": 1077
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3515,
      "step": 1078
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.4167,
      "step": 1079
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3007,
      "step": 1080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3505,
      "step": 1081
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.1987,
      "step": 1082
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3311,
      "step": 1083
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3327,
      "step": 1084
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.4333,
      "step": 1085
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4852,
      "step": 1086
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3692,
      "step": 1087
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2804,
      "step": 1088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2757,
      "step": 1089
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2836,
      "step": 1090
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.36,
      "step": 1091
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3251,
      "step": 1092
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2676,
      "step": 1093
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3695,
      "step": 1094
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3958,
      "step": 1095
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2812,
      "step": 1096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3103,
      "step": 1097
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3681,
      "step": 1098
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3071,
      "step": 1099
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4287,
      "step": 1100
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3983,
      "step": 1101
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3019,
      "step": 1102
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3475,
      "step": 1103
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3823,
      "step": 1104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.371,
      "step": 1105
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2103,
      "step": 1106
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3499,
      "step": 1107
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3728,
      "step": 1108
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2749,
      "step": 1109
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3678,
      "step": 1110
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.444,
      "step": 1111
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3431,
      "step": 1112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3941,
      "step": 1113
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.335,
      "step": 1114
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2307,
      "step": 1115
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4715,
      "step": 1116
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3797,
      "step": 1117
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3209,
      "step": 1118
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3698,
      "step": 1119
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2383,
      "step": 1120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3897,
      "step": 1121
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3363,
      "step": 1122
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4683,
      "step": 1123
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3432,
      "step": 1124
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4679,
      "step": 1125
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2691,
      "step": 1126
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2982,
      "step": 1127
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2467,
      "step": 1128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.2586,
      "step": 1129
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3816,
      "step": 1130
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4294,
      "step": 1131
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4064,
      "step": 1132
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4467,
      "step": 1133
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2648,
      "step": 1134
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2737,
      "step": 1135
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2972,
      "step": 1136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3052,
      "step": 1137
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2764,
      "step": 1138
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2214,
      "step": 1139
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2559,
      "step": 1140
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2989,
      "step": 1141
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3718,
      "step": 1142
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3936,
      "step": 1143
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3278,
      "step": 1144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2424,
      "step": 1145
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.291,
      "step": 1146
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2476,
      "step": 1147
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3314,
      "step": 1148
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.4345,
      "step": 1149
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2737,
      "step": 1150
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3019,
      "step": 1151
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3984,
      "step": 1152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.4365,
      "step": 1153
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.4559,
      "step": 1154
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.33,
      "step": 1155
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2701,
      "step": 1156
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3531,
      "step": 1157
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.459,
      "step": 1158
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.1893,
      "step": 1159
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3033,
      "step": 1160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3243,
      "step": 1161
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2844,
      "step": 1162
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.4774,
      "step": 1163
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.361,
      "step": 1164
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.291,
      "step": 1165
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2967,
      "step": 1166
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3633,
      "step": 1167
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2392,
      "step": 1168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3356,
      "step": 1169
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2664,
      "step": 1170
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2522,
      "step": 1171
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3911,
      "step": 1172
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3153,
      "step": 1173
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3247,
      "step": 1174
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3252,
      "step": 1175
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3449,
      "step": 1176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2576,
      "step": 1177
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2894,
      "step": 1178
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3495,
      "step": 1179
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3257,
      "step": 1180
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.251,
      "step": 1181
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.2906,
      "step": 1182
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.362,
      "step": 1183
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3473,
      "step": 1184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5011,
      "step": 1185
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2469,
      "step": 1186
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2893,
      "step": 1187
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3803,
      "step": 1188
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.274,
      "step": 1189
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2162,
      "step": 1190
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.311,
      "step": 1191
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4521,
      "step": 1192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3156,
      "step": 1193
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2632,
      "step": 1194
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3344,
      "step": 1195
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5234,
      "step": 1196
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.227,
      "step": 1197
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3469,
      "step": 1198
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3397,
      "step": 1199
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2175,
      "step": 1200
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.3368752598762512,
      "eval_runtime": 159.297,
      "eval_samples_per_second": 6.278,
      "eval_steps_per_second": 3.139,
      "step": 1200
    },
    {
      "epoch": 0.25,
      "mmlu_eval_accuracy": 0.604796655596056,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.8125,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6551724137931034,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.5454545454545454,
      "mmlu_eval_accuracy_college_medicine": 0.6818181818181818,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.5769230769230769,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.75,
      "mmlu_eval_accuracy_elementary_mathematics": 0.43902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.625,
      "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": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.627906976744186,
      "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": 0.8833333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7558139534883721,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.32,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.5142857142857142,
      "mmlu_eval_accuracy_professional_accounting": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_law": 0.4,
      "mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6666666666666666,
      "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": 0.8421052631578947,
      "mmlu_loss": 1.4258578755447822,
      "step": 1200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3934,
      "step": 1201
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2435,
      "step": 1202
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3464,
      "step": 1203
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3642,
      "step": 1204
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3491,
      "step": 1205
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2644,
      "step": 1206
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2944,
      "step": 1207
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3238,
      "step": 1208
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2265,
      "step": 1209
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2891,
      "step": 1210
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3696,
      "step": 1211
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3022,
      "step": 1212
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2857,
      "step": 1213
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4168,
      "step": 1214
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3136,
      "step": 1215
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.363,
      "step": 1216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.345,
      "step": 1217
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3945,
      "step": 1218
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2759,
      "step": 1219
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.297,
      "step": 1220
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.237,
      "step": 1221
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6326,
      "step": 1222
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2829,
      "step": 1223
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.232,
      "step": 1224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.2952,
      "step": 1225
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3061,
      "step": 1226
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3009,
      "step": 1227
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3351,
      "step": 1228
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3436,
      "step": 1229
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3163,
      "step": 1230
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3444,
      "step": 1231
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.2698,
      "step": 1232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.318,
      "step": 1233
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.2286,
      "step": 1234
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.344,
      "step": 1235
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3274,
      "step": 1236
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.2399,
      "step": 1237
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.299,
      "step": 1238
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5281,
      "step": 1239
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.2902,
      "step": 1240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3541,
      "step": 1241
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3137,
      "step": 1242
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4756,
      "step": 1243
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3564,
      "step": 1244
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4608,
      "step": 1245
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.438,
      "step": 1246
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4233,
      "step": 1247
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4087,
      "step": 1248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4123,
      "step": 1249
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3395,
      "step": 1250
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3677,
      "step": 1251
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3982,
      "step": 1252
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4741,
      "step": 1253
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3421,
      "step": 1254
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3133,
      "step": 1255
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3753,
      "step": 1256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.2696,
      "step": 1257
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3316,
      "step": 1258
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4674,
      "step": 1259
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.2973,
      "step": 1260
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.2093,
      "step": 1261
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3269,
      "step": 1262
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4071,
      "step": 1263
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4745,
      "step": 1264
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3916,
      "step": 1265
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.2964,
      "step": 1266
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4568,
      "step": 1267
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.2627,
      "step": 1268
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3151,
      "step": 1269
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.2679,
      "step": 1270
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3435,
      "step": 1271
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3807,
      "step": 1272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.2732,
      "step": 1273
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.353,
      "step": 1274
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.404,
      "step": 1275
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.304,
      "step": 1276
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4008,
      "step": 1277
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.2743,
      "step": 1278
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3276,
      "step": 1279
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3674,
      "step": 1280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2979,
      "step": 1281
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2657,
      "step": 1282
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.314,
      "step": 1283
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.293,
      "step": 1284
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4689,
      "step": 1285
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2089,
      "step": 1286
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3878,
      "step": 1287
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3381,
      "step": 1288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.377,
      "step": 1289
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.319,
      "step": 1290
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3792,
      "step": 1291
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2516,
      "step": 1292
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4335,
      "step": 1293
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.384,
      "step": 1294
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3692,
      "step": 1295
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2112,
      "step": 1296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3993,
      "step": 1297
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4006,
      "step": 1298
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2382,
      "step": 1299
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3246,
      "step": 1300
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3233,
      "step": 1301
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4042,
      "step": 1302
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.313,
      "step": 1303
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2501,
      "step": 1304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2538,
      "step": 1305
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3374,
      "step": 1306
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2801,
      "step": 1307
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.344,
      "step": 1308
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.446,
      "step": 1309
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4066,
      "step": 1310
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2646,
      "step": 1311
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2941,
      "step": 1312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4165,
      "step": 1313
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2693,
      "step": 1314
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3053,
      "step": 1315
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3058,
      "step": 1316
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2666,
      "step": 1317
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3697,
      "step": 1318
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3374,
      "step": 1319
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4698,
      "step": 1320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2904,
      "step": 1321
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3597,
      "step": 1322
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2734,
      "step": 1323
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.287,
      "step": 1324
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.2877,
      "step": 1325
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.3607,
      "step": 1326
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.28,
      "step": 1327
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2711,
      "step": 1328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4512,
      "step": 1329
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2502,
      "step": 1330
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4236,
      "step": 1331
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.48,
      "step": 1332
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4059,
      "step": 1333
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4075,
      "step": 1334
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2909,
      "step": 1335
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3353,
      "step": 1336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2889,
      "step": 1337
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2324,
      "step": 1338
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2577,
      "step": 1339
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.309,
      "step": 1340
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2902,
      "step": 1341
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3583,
      "step": 1342
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3089,
      "step": 1343
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3162,
      "step": 1344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3177,
      "step": 1345
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3074,
      "step": 1346
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.337,
      "step": 1347
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3847,
      "step": 1348
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3894,
      "step": 1349
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2633,
      "step": 1350
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2988,
      "step": 1351
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3381,
      "step": 1352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.331,
      "step": 1353
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3434,
      "step": 1354
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3924,
      "step": 1355
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4081,
      "step": 1356
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.297,
      "step": 1357
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.373,
      "step": 1358
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.329,
      "step": 1359
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3728,
      "step": 1360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3213,
      "step": 1361
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3627,
      "step": 1362
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3423,
      "step": 1363
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2909,
      "step": 1364
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2971,
      "step": 1365
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.3209,
      "step": 1366
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.336,
      "step": 1367
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4604,
      "step": 1368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6131,
      "step": 1369
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2421,
      "step": 1370
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2734,
      "step": 1371
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.283,
      "step": 1372
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.352,
      "step": 1373
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4599,
      "step": 1374
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.2357,
      "step": 1375
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2923,
      "step": 1376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2678,
      "step": 1377
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5359,
      "step": 1378
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.291,
      "step": 1379
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3565,
      "step": 1380
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4448,
      "step": 1381
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3475,
      "step": 1382
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3464,
      "step": 1383
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3087,
      "step": 1384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2698,
      "step": 1385
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3618,
      "step": 1386
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2829,
      "step": 1387
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.272,
      "step": 1388
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.305,
      "step": 1389
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3789,
      "step": 1390
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2664,
      "step": 1391
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2509,
      "step": 1392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2936,
      "step": 1393
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3184,
      "step": 1394
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3389,
      "step": 1395
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2824,
      "step": 1396
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.343,
      "step": 1397
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3114,
      "step": 1398
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4532,
      "step": 1399
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3418,
      "step": 1400
    },
    {
      "epoch": 0.29,
      "eval_loss": 0.33663323521614075,
      "eval_runtime": 159.4231,
      "eval_samples_per_second": 6.273,
      "eval_steps_per_second": 3.136,
      "step": 1400
    },
    {
      "epoch": 0.29,
      "mmlu_eval_accuracy": 0.5988431625993802,
      "mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
      "mmlu_eval_accuracy_anatomy": 0.2857142857142857,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "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": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.6538461538461539,
      "mmlu_eval_accuracy_econometrics": 0.6666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.5625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.5365853658536586,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "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": 0.7222222222222222,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5348837209302325,
      "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": 0.85,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "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": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.34,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.4857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.5806451612903226,
      "mmlu_eval_accuracy_professional_law": 0.4,
      "mmlu_eval_accuracy_professional_medicine": 0.6774193548387096,
      "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": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.6111111111111112,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.4995008165079993,
      "step": 1400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2955,
      "step": 1401
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2544,
      "step": 1402
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4088,
      "step": 1403
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2959,
      "step": 1404
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4989,
      "step": 1405
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6567,
      "step": 1406
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2983,
      "step": 1407
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2839,
      "step": 1408
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3097,
      "step": 1409
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4868,
      "step": 1410
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4287,
      "step": 1411
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2474,
      "step": 1412
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.335,
      "step": 1413
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2997,
      "step": 1414
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3079,
      "step": 1415
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2379,
      "step": 1416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2006,
      "step": 1417
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3114,
      "step": 1418
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3807,
      "step": 1419
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3207,
      "step": 1420
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3256,
      "step": 1421
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.2834,
      "step": 1422
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6884,
      "step": 1423
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.2604,
      "step": 1424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3474,
      "step": 1425
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3066,
      "step": 1426
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4209,
      "step": 1427
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4044,
      "step": 1428
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5416,
      "step": 1429
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.2346,
      "step": 1430
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4765,
      "step": 1431
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.363,
      "step": 1432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3189,
      "step": 1433
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.2185,
      "step": 1434
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3477,
      "step": 1435
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.2465,
      "step": 1436
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.276,
      "step": 1437
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3196,
      "step": 1438
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5027,
      "step": 1439
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3729,
      "step": 1440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3676,
      "step": 1441
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.2836,
      "step": 1442
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.2014,
      "step": 1443
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3293,
      "step": 1444
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6591,
      "step": 1445
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.2323,
      "step": 1446
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3745,
      "step": 1447
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7213,
      "step": 1448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3208,
      "step": 1449
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.2419,
      "step": 1450
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3249,
      "step": 1451
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.2996,
      "step": 1452
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3522,
      "step": 1453
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3228,
      "step": 1454
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3358,
      "step": 1455
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4661,
      "step": 1456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.361,
      "step": 1457
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3583,
      "step": 1458
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.301,
      "step": 1459
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3003,
      "step": 1460
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.2422,
      "step": 1461
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5315,
      "step": 1462
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3418,
      "step": 1463
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3769,
      "step": 1464
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.2861,
      "step": 1465
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.2097,
      "step": 1466
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.336,
      "step": 1467
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3407,
      "step": 1468
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3419,
      "step": 1469
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3095,
      "step": 1470
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.3398,
      "step": 1471
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.283,
      "step": 1472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4463,
      "step": 1473
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2983,
      "step": 1474
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2905,
      "step": 1475
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2503,
      "step": 1476
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4403,
      "step": 1477
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3506,
      "step": 1478
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3436,
      "step": 1479
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3225,
      "step": 1480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2525,
      "step": 1481
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2677,
      "step": 1482
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4132,
      "step": 1483
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2935,
      "step": 1484
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3177,
      "step": 1485
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4628,
      "step": 1486
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3345,
      "step": 1487
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2949,
      "step": 1488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2881,
      "step": 1489
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3309,
      "step": 1490
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.358,
      "step": 1491
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.291,
      "step": 1492
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2234,
      "step": 1493
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3209,
      "step": 1494
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3667,
      "step": 1495
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4149,
      "step": 1496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3423,
      "step": 1497
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3355,
      "step": 1498
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.326,
      "step": 1499
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.301,
      "step": 1500
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4943,
      "step": 1501
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3528,
      "step": 1502
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2953,
      "step": 1503
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3267,
      "step": 1504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2319,
      "step": 1505
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2877,
      "step": 1506
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2859,
      "step": 1507
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4401,
      "step": 1508
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4634,
      "step": 1509
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.247,
      "step": 1510
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.377,
      "step": 1511
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2854,
      "step": 1512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.308,
      "step": 1513
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3239,
      "step": 1514
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3237,
      "step": 1515
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3177,
      "step": 1516
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2864,
      "step": 1517
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.2372,
      "step": 1518
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3515,
      "step": 1519
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2747,
      "step": 1520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2489,
      "step": 1521
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4183,
      "step": 1522
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4211,
      "step": 1523
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3316,
      "step": 1524
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3223,
      "step": 1525
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2908,
      "step": 1526
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5003,
      "step": 1527
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3043,
      "step": 1528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4197,
      "step": 1529
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3874,
      "step": 1530
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3501,
      "step": 1531
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3674,
      "step": 1532
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2134,
      "step": 1533
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.364,
      "step": 1534
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2458,
      "step": 1535
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2149,
      "step": 1536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3037,
      "step": 1537
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3675,
      "step": 1538
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2494,
      "step": 1539
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4273,
      "step": 1540
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3962,
      "step": 1541
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3332,
      "step": 1542
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3336,
      "step": 1543
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3131,
      "step": 1544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3025,
      "step": 1545
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2949,
      "step": 1546
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4005,
      "step": 1547
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.258,
      "step": 1548
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2922,
      "step": 1549
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2556,
      "step": 1550
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2101,
      "step": 1551
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2162,
      "step": 1552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.337,
      "step": 1553
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3086,
      "step": 1554
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4988,
      "step": 1555
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3734,
      "step": 1556
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4444,
      "step": 1557
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4434,
      "step": 1558
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2584,
      "step": 1559
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.1929,
      "step": 1560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3251,
      "step": 1561
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.209,
      "step": 1562
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2978,
      "step": 1563
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2986,
      "step": 1564
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.41,
      "step": 1565
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.394,
      "step": 1566
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.2986,
      "step": 1567
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.3496,
      "step": 1568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3498,
      "step": 1569
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3109,
      "step": 1570
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2075,
      "step": 1571
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3613,
      "step": 1572
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2628,
      "step": 1573
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3275,
      "step": 1574
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2889,
      "step": 1575
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3382,
      "step": 1576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2797,
      "step": 1577
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5451,
      "step": 1578
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3645,
      "step": 1579
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.264,
      "step": 1580
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3261,
      "step": 1581
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2957,
      "step": 1582
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2267,
      "step": 1583
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2626,
      "step": 1584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3648,
      "step": 1585
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5535,
      "step": 1586
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3266,
      "step": 1587
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3017,
      "step": 1588
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2288,
      "step": 1589
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3663,
      "step": 1590
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2932,
      "step": 1591
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3949,
      "step": 1592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2377,
      "step": 1593
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3316,
      "step": 1594
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3649,
      "step": 1595
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3073,
      "step": 1596
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3401,
      "step": 1597
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3836,
      "step": 1598
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2566,
      "step": 1599
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4909,
      "step": 1600
    },
    {
      "epoch": 0.33,
      "eval_loss": 0.3317873179912567,
      "eval_runtime": 159.3941,
      "eval_samples_per_second": 6.274,
      "eval_steps_per_second": 3.137,
      "step": 1600
    },
    {
      "epoch": 0.33,
      "mmlu_eval_accuracy": 0.604104053743682,
      "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": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.6818181818181818,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.6875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.5609756097560976,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.53125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.627906976744186,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3448275862068966,
      "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": 0.6923076923076923,
      "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": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7790697674418605,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.33,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.5142857142857142,
      "mmlu_eval_accuracy_professional_accounting": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_law": 0.40588235294117647,
      "mmlu_eval_accuracy_professional_medicine": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.9545454545454546,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.6666666666666666,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.4852339224622393,
      "step": 1600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3334,
      "step": 1601
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4838,
      "step": 1602
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3516,
      "step": 1603
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.25,
      "step": 1604
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.49,
      "step": 1605
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2733,
      "step": 1606
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2783,
      "step": 1607
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2539,
      "step": 1608
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3256,
      "step": 1609
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3972,
      "step": 1610
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4074,
      "step": 1611
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.379,
      "step": 1612
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3495,
      "step": 1613
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.3087,
      "step": 1614
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2345,
      "step": 1615
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.2718,
      "step": 1616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3027,
      "step": 1617
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2855,
      "step": 1618
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3173,
      "step": 1619
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2171,
      "step": 1620
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2817,
      "step": 1621
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3155,
      "step": 1622
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4401,
      "step": 1623
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4339,
      "step": 1624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3175,
      "step": 1625
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2457,
      "step": 1626
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.269,
      "step": 1627
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3708,
      "step": 1628
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2432,
      "step": 1629
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2381,
      "step": 1630
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3507,
      "step": 1631
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2882,
      "step": 1632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2381,
      "step": 1633
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3296,
      "step": 1634
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3083,
      "step": 1635
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4051,
      "step": 1636
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3996,
      "step": 1637
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2851,
      "step": 1638
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3441,
      "step": 1639
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4172,
      "step": 1640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3062,
      "step": 1641
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3192,
      "step": 1642
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2048,
      "step": 1643
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3372,
      "step": 1644
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2578,
      "step": 1645
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3007,
      "step": 1646
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3709,
      "step": 1647
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3193,
      "step": 1648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3146,
      "step": 1649
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2601,
      "step": 1650
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4825,
      "step": 1651
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2891,
      "step": 1652
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2325,
      "step": 1653
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3162,
      "step": 1654
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2901,
      "step": 1655
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4135,
      "step": 1656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3579,
      "step": 1657
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2688,
      "step": 1658
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2458,
      "step": 1659
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4185,
      "step": 1660
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3543,
      "step": 1661
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3159,
      "step": 1662
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.2966,
      "step": 1663
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3614,
      "step": 1664
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2745,
      "step": 1665
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3058,
      "step": 1666
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4681,
      "step": 1667
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.328,
      "step": 1668
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2062,
      "step": 1669
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4278,
      "step": 1670
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4948,
      "step": 1671
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.278,
      "step": 1672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3436,
      "step": 1673
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3287,
      "step": 1674
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3692,
      "step": 1675
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5099,
      "step": 1676
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3628,
      "step": 1677
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2621,
      "step": 1678
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3419,
      "step": 1679
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.32,
      "step": 1680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.233,
      "step": 1681
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2551,
      "step": 1682
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2684,
      "step": 1683
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2937,
      "step": 1684
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3269,
      "step": 1685
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2774,
      "step": 1686
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.345,
      "step": 1687
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3396,
      "step": 1688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2436,
      "step": 1689
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.329,
      "step": 1690
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2881,
      "step": 1691
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4038,
      "step": 1692
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2867,
      "step": 1693
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2699,
      "step": 1694
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5154,
      "step": 1695
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3967,
      "step": 1696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2934,
      "step": 1697
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2944,
      "step": 1698
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2675,
      "step": 1699
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2601,
      "step": 1700
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.347,
      "step": 1701
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4217,
      "step": 1702
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2517,
      "step": 1703
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4265,
      "step": 1704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4762,
      "step": 1705
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.442,
      "step": 1706
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3528,
      "step": 1707
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2042,
      "step": 1708
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3589,
      "step": 1709
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3871,
      "step": 1710
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.3317,
      "step": 1711
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.2544,
      "step": 1712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4393,
      "step": 1713
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2329,
      "step": 1714
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5222,
      "step": 1715
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.388,
      "step": 1716
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3759,
      "step": 1717
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4002,
      "step": 1718
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3111,
      "step": 1719
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4526,
      "step": 1720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3783,
      "step": 1721
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4324,
      "step": 1722
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3174,
      "step": 1723
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2132,
      "step": 1724
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2844,
      "step": 1725
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2983,
      "step": 1726
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2927,
      "step": 1727
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2667,
      "step": 1728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2995,
      "step": 1729
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3512,
      "step": 1730
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2147,
      "step": 1731
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2851,
      "step": 1732
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2828,
      "step": 1733
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3263,
      "step": 1734
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3379,
      "step": 1735
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3372,
      "step": 1736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2135,
      "step": 1737
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3782,
      "step": 1738
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3953,
      "step": 1739
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2299,
      "step": 1740
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2322,
      "step": 1741
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3818,
      "step": 1742
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3403,
      "step": 1743
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3873,
      "step": 1744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2827,
      "step": 1745
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2546,
      "step": 1746
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3017,
      "step": 1747
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4125,
      "step": 1748
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2431,
      "step": 1749
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3023,
      "step": 1750
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5773,
      "step": 1751
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3329,
      "step": 1752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3149,
      "step": 1753
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.257,
      "step": 1754
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2559,
      "step": 1755
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2261,
      "step": 1756
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.285,
      "step": 1757
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2734,
      "step": 1758
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.296,
      "step": 1759
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.2676,
      "step": 1760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.3249,
      "step": 1761
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2878,
      "step": 1762
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2918,
      "step": 1763
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2361,
      "step": 1764
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2812,
      "step": 1765
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4694,
      "step": 1766
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3372,
      "step": 1767
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.371,
      "step": 1768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4062,
      "step": 1769
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4353,
      "step": 1770
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4231,
      "step": 1771
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3154,
      "step": 1772
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3907,
      "step": 1773
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3659,
      "step": 1774
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2288,
      "step": 1775
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3506,
      "step": 1776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2445,
      "step": 1777
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3384,
      "step": 1778
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3252,
      "step": 1779
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3783,
      "step": 1780
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2369,
      "step": 1781
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2717,
      "step": 1782
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2256,
      "step": 1783
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2701,
      "step": 1784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2927,
      "step": 1785
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3099,
      "step": 1786
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3354,
      "step": 1787
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6063,
      "step": 1788
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3382,
      "step": 1789
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.375,
      "step": 1790
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2494,
      "step": 1791
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3243,
      "step": 1792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3112,
      "step": 1793
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3974,
      "step": 1794
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3529,
      "step": 1795
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2518,
      "step": 1796
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6142,
      "step": 1797
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.266,
      "step": 1798
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3405,
      "step": 1799
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4709,
      "step": 1800
    },
    {
      "epoch": 0.37,
      "eval_loss": 0.3296074867248535,
      "eval_runtime": 159.4124,
      "eval_samples_per_second": 6.273,
      "eval_steps_per_second": 3.137,
      "step": 1800
    },
    {
      "epoch": 0.37,
      "mmlu_eval_accuracy": 0.5929397997737007,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "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": 0.5,
      "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": 0.7272727272727273,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.6875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4878048780487805,
      "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": 0.5555555555555556,
      "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": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5769230769230769,
      "mmlu_eval_accuracy_high_school_physics": 0.058823529411764705,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "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": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7558139534883721,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.37,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.4857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_law": 0.40588235294117647,
      "mmlu_eval_accuracy_professional_medicine": 0.7096774193548387,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "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": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.290066049864647,
      "step": 1800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.476,
      "step": 1801
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4168,
      "step": 1802
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4114,
      "step": 1803
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.313,
      "step": 1804
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4064,
      "step": 1805
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2534,
      "step": 1806
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.2692,
      "step": 1807
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.3223,
      "step": 1808
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.236,
      "step": 1809
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2488,
      "step": 1810
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3263,
      "step": 1811
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3254,
      "step": 1812
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3205,
      "step": 1813
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4668,
      "step": 1814
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2813,
      "step": 1815
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3845,
      "step": 1816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3452,
      "step": 1817
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2665,
      "step": 1818
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3389,
      "step": 1819
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3818,
      "step": 1820
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2378,
      "step": 1821
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2967,
      "step": 1822
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5581,
      "step": 1823
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2934,
      "step": 1824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2547,
      "step": 1825
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4571,
      "step": 1826
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2484,
      "step": 1827
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5326,
      "step": 1828
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4539,
      "step": 1829
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3168,
      "step": 1830
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4649,
      "step": 1831
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2276,
      "step": 1832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.296,
      "step": 1833
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2417,
      "step": 1834
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4738,
      "step": 1835
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2447,
      "step": 1836
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.344,
      "step": 1837
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4444,
      "step": 1838
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.336,
      "step": 1839
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2506,
      "step": 1840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.316,
      "step": 1841
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3005,
      "step": 1842
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.295,
      "step": 1843
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3407,
      "step": 1844
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3985,
      "step": 1845
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3267,
      "step": 1846
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3497,
      "step": 1847
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2429,
      "step": 1848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4069,
      "step": 1849
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2855,
      "step": 1850
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3132,
      "step": 1851
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2429,
      "step": 1852
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3009,
      "step": 1853
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.256,
      "step": 1854
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2679,
      "step": 1855
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3141,
      "step": 1856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.2826,
      "step": 1857
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3554,
      "step": 1858
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.2881,
      "step": 1859
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3123,
      "step": 1860
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3731,
      "step": 1861
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4021,
      "step": 1862
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.288,
      "step": 1863
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3879,
      "step": 1864
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.2616,
      "step": 1865
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3468,
      "step": 1866
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3225,
      "step": 1867
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3458,
      "step": 1868
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4093,
      "step": 1869
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4435,
      "step": 1870
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3674,
      "step": 1871
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3387,
      "step": 1872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.2799,
      "step": 1873
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.2873,
      "step": 1874
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3178,
      "step": 1875
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.2839,
      "step": 1876
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3371,
      "step": 1877
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3523,
      "step": 1878
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.258,
      "step": 1879
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3013,
      "step": 1880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.2597,
      "step": 1881
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4445,
      "step": 1882
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3611,
      "step": 1883
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3538,
      "step": 1884
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3116,
      "step": 1885
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3309,
      "step": 1886
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4336,
      "step": 1887
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3187,
      "step": 1888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4095,
      "step": 1889
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3019,
      "step": 1890
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3199,
      "step": 1891
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3413,
      "step": 1892
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3332,
      "step": 1893
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3967,
      "step": 1894
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3117,
      "step": 1895
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.2655,
      "step": 1896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5061,
      "step": 1897
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.2549,
      "step": 1898
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3285,
      "step": 1899
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3267,
      "step": 1900
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.2332,
      "step": 1901
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.2257,
      "step": 1902
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.2173,
      "step": 1903
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.297,
      "step": 1904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.3259,
      "step": 1905
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3748,
      "step": 1906
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3784,
      "step": 1907
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2456,
      "step": 1908
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2993,
      "step": 1909
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3422,
      "step": 1910
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4405,
      "step": 1911
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3775,
      "step": 1912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2902,
      "step": 1913
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2634,
      "step": 1914
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3973,
      "step": 1915
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3364,
      "step": 1916
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3295,
      "step": 1917
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2572,
      "step": 1918
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4997,
      "step": 1919
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2747,
      "step": 1920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4602,
      "step": 1921
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4479,
      "step": 1922
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3899,
      "step": 1923
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2504,
      "step": 1924
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.1924,
      "step": 1925
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.371,
      "step": 1926
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3654,
      "step": 1927
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2943,
      "step": 1928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2598,
      "step": 1929
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4397,
      "step": 1930
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4327,
      "step": 1931
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3373,
      "step": 1932
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6977,
      "step": 1933
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2642,
      "step": 1934
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3059,
      "step": 1935
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3988,
      "step": 1936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3383,
      "step": 1937
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3194,
      "step": 1938
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3463,
      "step": 1939
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.387,
      "step": 1940
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2404,
      "step": 1941
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3116,
      "step": 1942
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.1949,
      "step": 1943
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2763,
      "step": 1944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2538,
      "step": 1945
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.314,
      "step": 1946
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3899,
      "step": 1947
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2852,
      "step": 1948
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.2112,
      "step": 1949
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4189,
      "step": 1950
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3179,
      "step": 1951
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.1909,
      "step": 1952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.324,
      "step": 1953
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.3513,
      "step": 1954
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.2538,
      "step": 1955
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.2658,
      "step": 1956
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.2369,
      "step": 1957
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3682,
      "step": 1958
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.34,
      "step": 1959
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3278,
      "step": 1960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4265,
      "step": 1961
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.2768,
      "step": 1962
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4373,
      "step": 1963
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3404,
      "step": 1964
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.31,
      "step": 1965
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.2805,
      "step": 1966
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.307,
      "step": 1967
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3091,
      "step": 1968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.2087,
      "step": 1969
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4403,
      "step": 1970
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4126,
      "step": 1971
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.324,
      "step": 1972
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.1901,
      "step": 1973
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3296,
      "step": 1974
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3338,
      "step": 1975
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3299,
      "step": 1976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.299,
      "step": 1977
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3758,
      "step": 1978
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3832,
      "step": 1979
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.239,
      "step": 1980
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3494,
      "step": 1981
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3546,
      "step": 1982
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.215,
      "step": 1983
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.2356,
      "step": 1984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3154,
      "step": 1985
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3424,
      "step": 1986
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3361,
      "step": 1987
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3563,
      "step": 1988
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.2955,
      "step": 1989
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3878,
      "step": 1990
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.2491,
      "step": 1991
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3823,
      "step": 1992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3914,
      "step": 1993
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3052,
      "step": 1994
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.2639,
      "step": 1995
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3143,
      "step": 1996
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3359,
      "step": 1997
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.2186,
      "step": 1998
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3563,
      "step": 1999
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.2322,
      "step": 2000
    },
    {
      "epoch": 0.41,
      "eval_loss": 0.3272154629230499,
      "eval_runtime": 159.4111,
      "eval_samples_per_second": 6.273,
      "eval_steps_per_second": 3.137,
      "step": 2000
    },
    {
      "epoch": 0.41,
      "mmlu_eval_accuracy": 0.5934577579656638,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "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": 0.5454545454545454,
      "mmlu_eval_accuracy_college_medicine": 0.7272727272727273,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.5,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.07142857142857142,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.53125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.6046511627906976,
      "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "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": 0.7272727272727273,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7790697674418605,
      "mmlu_eval_accuracy_moral_disputes": 0.6052631578947368,
      "mmlu_eval_accuracy_moral_scenarios": 0.3,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.4857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_law": 0.3588235294117647,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "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": 0.8421052631578947,
      "mmlu_loss": 1.3003345157746857,
      "step": 2000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3727,
      "step": 2001
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.3079,
      "step": 2002
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2573,
      "step": 2003
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3337,
      "step": 2004
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2911,
      "step": 2005
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2821,
      "step": 2006
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4295,
      "step": 2007
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4014,
      "step": 2008
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2357,
      "step": 2009
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4374,
      "step": 2010
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2832,
      "step": 2011
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2465,
      "step": 2012
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3585,
      "step": 2013
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3437,
      "step": 2014
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3483,
      "step": 2015
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6036,
      "step": 2016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2453,
      "step": 2017
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4562,
      "step": 2018
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2942,
      "step": 2019
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2823,
      "step": 2020
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4462,
      "step": 2021
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4502,
      "step": 2022
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3407,
      "step": 2023
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2783,
      "step": 2024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3865,
      "step": 2025
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2058,
      "step": 2026
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3276,
      "step": 2027
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2036,
      "step": 2028
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4829,
      "step": 2029
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4307,
      "step": 2030
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4514,
      "step": 2031
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3514,
      "step": 2032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2469,
      "step": 2033
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3152,
      "step": 2034
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.224,
      "step": 2035
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2342,
      "step": 2036
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3168,
      "step": 2037
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2909,
      "step": 2038
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2635,
      "step": 2039
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4834,
      "step": 2040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3541,
      "step": 2041
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5334,
      "step": 2042
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3128,
      "step": 2043
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4182,
      "step": 2044
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3253,
      "step": 2045
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2585,
      "step": 2046
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2846,
      "step": 2047
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3583,
      "step": 2048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2846,
      "step": 2049
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.2048,
      "step": 2050
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2797,
      "step": 2051
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2851,
      "step": 2052
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2996,
      "step": 2053
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4072,
      "step": 2054
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3137,
      "step": 2055
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4602,
      "step": 2056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2964,
      "step": 2057
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3743,
      "step": 2058
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.365,
      "step": 2059
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4759,
      "step": 2060
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3202,
      "step": 2061
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2444,
      "step": 2062
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2957,
      "step": 2063
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3834,
      "step": 2064
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2627,
      "step": 2065
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3752,
      "step": 2066
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2684,
      "step": 2067
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3806,
      "step": 2068
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2847,
      "step": 2069
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2859,
      "step": 2070
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2757,
      "step": 2071
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3033,
      "step": 2072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2415,
      "step": 2073
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3397,
      "step": 2074
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3922,
      "step": 2075
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2845,
      "step": 2076
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3283,
      "step": 2077
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3798,
      "step": 2078
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2773,
      "step": 2079
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.447,
      "step": 2080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.1736,
      "step": 2081
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3346,
      "step": 2082
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3686,
      "step": 2083
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2923,
      "step": 2084
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2911,
      "step": 2085
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3452,
      "step": 2086
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2739,
      "step": 2087
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2816,
      "step": 2088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2528,
      "step": 2089
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4857,
      "step": 2090
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2726,
      "step": 2091
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3006,
      "step": 2092
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4625,
      "step": 2093
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3913,
      "step": 2094
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.3544,
      "step": 2095
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4483,
      "step": 2096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2367,
      "step": 2097
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.2689,
      "step": 2098
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2727,
      "step": 2099
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.225,
      "step": 2100
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3641,
      "step": 2101
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4111,
      "step": 2102
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2746,
      "step": 2103
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3265,
      "step": 2104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3336,
      "step": 2105
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.342,
      "step": 2106
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2971,
      "step": 2107
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2752,
      "step": 2108
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2553,
      "step": 2109
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2752,
      "step": 2110
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4278,
      "step": 2111
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6237,
      "step": 2112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3716,
      "step": 2113
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3208,
      "step": 2114
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3103,
      "step": 2115
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2364,
      "step": 2116
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2568,
      "step": 2117
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2618,
      "step": 2118
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4231,
      "step": 2119
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3675,
      "step": 2120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2289,
      "step": 2121
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2614,
      "step": 2122
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2486,
      "step": 2123
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3033,
      "step": 2124
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3526,
      "step": 2125
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2326,
      "step": 2126
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.361,
      "step": 2127
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3362,
      "step": 2128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4127,
      "step": 2129
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3578,
      "step": 2130
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2288,
      "step": 2131
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2383,
      "step": 2132
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5872,
      "step": 2133
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2634,
      "step": 2134
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2918,
      "step": 2135
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4192,
      "step": 2136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2619,
      "step": 2137
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3554,
      "step": 2138
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3614,
      "step": 2139
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3947,
      "step": 2140
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3318,
      "step": 2141
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2404,
      "step": 2142
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2945,
      "step": 2143
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.3864,
      "step": 2144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2306,
      "step": 2145
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2789,
      "step": 2146
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.2943,
      "step": 2147
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2998,
      "step": 2148
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2925,
      "step": 2149
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3475,
      "step": 2150
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3818,
      "step": 2151
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.346,
      "step": 2152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3085,
      "step": 2153
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2997,
      "step": 2154
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2992,
      "step": 2155
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3433,
      "step": 2156
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2411,
      "step": 2157
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2752,
      "step": 2158
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2499,
      "step": 2159
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3008,
      "step": 2160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4845,
      "step": 2161
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3961,
      "step": 2162
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2607,
      "step": 2163
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4121,
      "step": 2164
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4364,
      "step": 2165
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3385,
      "step": 2166
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2857,
      "step": 2167
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3134,
      "step": 2168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2569,
      "step": 2169
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3019,
      "step": 2170
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.362,
      "step": 2171
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2862,
      "step": 2172
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2577,
      "step": 2173
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3548,
      "step": 2174
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3655,
      "step": 2175
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2628,
      "step": 2176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2398,
      "step": 2177
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2868,
      "step": 2178
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2844,
      "step": 2179
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4666,
      "step": 2180
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2177,
      "step": 2181
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2758,
      "step": 2182
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2587,
      "step": 2183
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3336,
      "step": 2184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5457,
      "step": 2185
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3178,
      "step": 2186
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3036,
      "step": 2187
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2872,
      "step": 2188
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3418,
      "step": 2189
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.3579,
      "step": 2190
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4109,
      "step": 2191
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5636,
      "step": 2192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2996,
      "step": 2193
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2677,
      "step": 2194
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.2698,
      "step": 2195
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.229,
      "step": 2196
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3256,
      "step": 2197
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3581,
      "step": 2198
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3056,
      "step": 2199
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4005,
      "step": 2200
    },
    {
      "epoch": 0.46,
      "eval_loss": 0.32537850737571716,
      "eval_runtime": 159.3549,
      "eval_samples_per_second": 6.275,
      "eval_steps_per_second": 3.138,
      "step": 2200
    },
    {
      "epoch": 0.46,
      "mmlu_eval_accuracy": 0.5984871069576992,
      "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": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "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": 0.7272727272727273,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.6875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
      "mmlu_eval_accuracy_formal_logic": 0.35714285714285715,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
      "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": 0.5769230769230769,
      "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": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "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": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7558139534883721,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.36,
      "mmlu_eval_accuracy_nutrition": 0.6060606060606061,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.5142857142857142,
      "mmlu_eval_accuracy_professional_accounting": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_law": 0.38823529411764707,
      "mmlu_eval_accuracy_professional_medicine": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_psychology": 0.6086956521739131,
      "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": 0.7272727272727273,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.3616631865190152,
      "step": 2200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3102,
      "step": 2201
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3595,
      "step": 2202
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.351,
      "step": 2203
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2891,
      "step": 2204
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3288,
      "step": 2205
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2724,
      "step": 2206
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2825,
      "step": 2207
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2631,
      "step": 2208
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2111,
      "step": 2209
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3155,
      "step": 2210
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2664,
      "step": 2211
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3148,
      "step": 2212
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3561,
      "step": 2213
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3888,
      "step": 2214
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3103,
      "step": 2215
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3748,
      "step": 2216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2563,
      "step": 2217
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3974,
      "step": 2218
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3893,
      "step": 2219
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2906,
      "step": 2220
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2931,
      "step": 2221
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2437,
      "step": 2222
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3034,
      "step": 2223
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.178,
      "step": 2224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5423,
      "step": 2225
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3217,
      "step": 2226
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3208,
      "step": 2227
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3174,
      "step": 2228
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3839,
      "step": 2229
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2959,
      "step": 2230
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2848,
      "step": 2231
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3395,
      "step": 2232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3169,
      "step": 2233
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2967,
      "step": 2234
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4553,
      "step": 2235
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3397,
      "step": 2236
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2678,
      "step": 2237
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.408,
      "step": 2238
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3016,
      "step": 2239
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4057,
      "step": 2240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3074,
      "step": 2241
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3394,
      "step": 2242
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.2467,
      "step": 2243
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3836,
      "step": 2244
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2899,
      "step": 2245
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.323,
      "step": 2246
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3371,
      "step": 2247
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3776,
      "step": 2248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3593,
      "step": 2249
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3213,
      "step": 2250
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3312,
      "step": 2251
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2718,
      "step": 2252
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2636,
      "step": 2253
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2744,
      "step": 2254
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3356,
      "step": 2255
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2204,
      "step": 2256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3204,
      "step": 2257
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4135,
      "step": 2258
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3245,
      "step": 2259
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4265,
      "step": 2260
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.402,
      "step": 2261
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3924,
      "step": 2262
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3205,
      "step": 2263
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2868,
      "step": 2264
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3436,
      "step": 2265
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2054,
      "step": 2266
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3007,
      "step": 2267
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3033,
      "step": 2268
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.243,
      "step": 2269
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3876,
      "step": 2270
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3135,
      "step": 2271
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2868,
      "step": 2272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2493,
      "step": 2273
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3066,
      "step": 2274
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2735,
      "step": 2275
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.337,
      "step": 2276
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2182,
      "step": 2277
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2997,
      "step": 2278
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2609,
      "step": 2279
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2795,
      "step": 2280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3076,
      "step": 2281
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2344,
      "step": 2282
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.6197,
      "step": 2283
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3671,
      "step": 2284
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2955,
      "step": 2285
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4142,
      "step": 2286
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3638,
      "step": 2287
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.2299,
      "step": 2288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5845,
      "step": 2289
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4752,
      "step": 2290
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.271,
      "step": 2291
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4283,
      "step": 2292
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3988,
      "step": 2293
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3024,
      "step": 2294
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3501,
      "step": 2295
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2533,
      "step": 2296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3485,
      "step": 2297
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2133,
      "step": 2298
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3762,
      "step": 2299
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3129,
      "step": 2300
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2974,
      "step": 2301
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2297,
      "step": 2302
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2267,
      "step": 2303
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2938,
      "step": 2304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3198,
      "step": 2305
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.365,
      "step": 2306
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4244,
      "step": 2307
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3538,
      "step": 2308
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5421,
      "step": 2309
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2354,
      "step": 2310
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3551,
      "step": 2311
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2621,
      "step": 2312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3262,
      "step": 2313
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3462,
      "step": 2314
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3521,
      "step": 2315
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2337,
      "step": 2316
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3702,
      "step": 2317
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2524,
      "step": 2318
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3518,
      "step": 2319
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2951,
      "step": 2320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2763,
      "step": 2321
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5593,
      "step": 2322
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.24,
      "step": 2323
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3651,
      "step": 2324
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3368,
      "step": 2325
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2324,
      "step": 2326
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4447,
      "step": 2327
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.311,
      "step": 2328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4286,
      "step": 2329
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3579,
      "step": 2330
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3178,
      "step": 2331
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3442,
      "step": 2332
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2438,
      "step": 2333
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3807,
      "step": 2334
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4419,
      "step": 2335
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2869,
      "step": 2336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.3095,
      "step": 2337
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4262,
      "step": 2338
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.2506,
      "step": 2339
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.587,
      "step": 2340
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3082,
      "step": 2341
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3208,
      "step": 2342
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2632,
      "step": 2343
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2658,
      "step": 2344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2921,
      "step": 2345
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3638,
      "step": 2346
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.396,
      "step": 2347
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4519,
      "step": 2348
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3211,
      "step": 2349
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3547,
      "step": 2350
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3231,
      "step": 2351
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2944,
      "step": 2352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2525,
      "step": 2353
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5334,
      "step": 2354
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3087,
      "step": 2355
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3054,
      "step": 2356
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3999,
      "step": 2357
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2628,
      "step": 2358
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2259,
      "step": 2359
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3732,
      "step": 2360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3282,
      "step": 2361
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4016,
      "step": 2362
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2494,
      "step": 2363
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2799,
      "step": 2364
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3576,
      "step": 2365
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2804,
      "step": 2366
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2849,
      "step": 2367
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.533,
      "step": 2368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3281,
      "step": 2369
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3889,
      "step": 2370
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2832,
      "step": 2371
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2694,
      "step": 2372
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3259,
      "step": 2373
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2757,
      "step": 2374
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3113,
      "step": 2375
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2023,
      "step": 2376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2795,
      "step": 2377
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3352,
      "step": 2378
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3902,
      "step": 2379
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3477,
      "step": 2380
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3441,
      "step": 2381
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3274,
      "step": 2382
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3105,
      "step": 2383
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4034,
      "step": 2384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.3221,
      "step": 2385
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4283,
      "step": 2386
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.2908,
      "step": 2387
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4211,
      "step": 2388
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3944,
      "step": 2389
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4042,
      "step": 2390
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.268,
      "step": 2391
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3898,
      "step": 2392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.285,
      "step": 2393
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4599,
      "step": 2394
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4067,
      "step": 2395
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.318,
      "step": 2396
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4227,
      "step": 2397
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4242,
      "step": 2398
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2925,
      "step": 2399
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.375,
      "step": 2400
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.32525983452796936,
      "eval_runtime": 159.4296,
      "eval_samples_per_second": 6.272,
      "eval_steps_per_second": 3.136,
      "step": 2400
    },
    {
      "epoch": 0.5,
      "mmlu_eval_accuracy": 0.5874332255577591,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.8125,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "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": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.6818181818181818,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "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": 0.43902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.07142857142857142,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.53125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5384615384615384,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8,
      "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": 0.8260869565217391,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "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": 0.7272727272727273,
      "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": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.38,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.5714285714285714,
      "mmlu_eval_accuracy_professional_accounting": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_law": 0.37058823529411766,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "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": 0.7272727272727273,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.4088074159139752,
      "step": 2400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2449,
      "step": 2401
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.258,
      "step": 2402
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3114,
      "step": 2403
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.1928,
      "step": 2404
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2217,
      "step": 2405
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2092,
      "step": 2406
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3191,
      "step": 2407
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3278,
      "step": 2408
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2455,
      "step": 2409
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2828,
      "step": 2410
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5289,
      "step": 2411
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2719,
      "step": 2412
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4659,
      "step": 2413
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.291,
      "step": 2414
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3285,
      "step": 2415
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3346,
      "step": 2416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3499,
      "step": 2417
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.333,
      "step": 2418
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3,
      "step": 2419
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3095,
      "step": 2420
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4378,
      "step": 2421
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3089,
      "step": 2422
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3421,
      "step": 2423
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3108,
      "step": 2424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2792,
      "step": 2425
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2811,
      "step": 2426
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3398,
      "step": 2427
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2785,
      "step": 2428
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2626,
      "step": 2429
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.245,
      "step": 2430
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2061,
      "step": 2431
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3246,
      "step": 2432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3268,
      "step": 2433
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.2439,
      "step": 2434
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3133,
      "step": 2435
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.3681,
      "step": 2436
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.288,
      "step": 2437
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2574,
      "step": 2438
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.272,
      "step": 2439
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2292,
      "step": 2440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.423,
      "step": 2441
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2919,
      "step": 2442
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5107,
      "step": 2443
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.253,
      "step": 2444
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3367,
      "step": 2445
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3684,
      "step": 2446
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2849,
      "step": 2447
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2894,
      "step": 2448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2311,
      "step": 2449
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3915,
      "step": 2450
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3682,
      "step": 2451
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3666,
      "step": 2452
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4559,
      "step": 2453
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2165,
      "step": 2454
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2365,
      "step": 2455
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3387,
      "step": 2456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2949,
      "step": 2457
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2424,
      "step": 2458
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3245,
      "step": 2459
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3115,
      "step": 2460
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3348,
      "step": 2461
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3116,
      "step": 2462
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4062,
      "step": 2463
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.402,
      "step": 2464
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3458,
      "step": 2465
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3407,
      "step": 2466
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3118,
      "step": 2467
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.256,
      "step": 2468
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3443,
      "step": 2469
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4113,
      "step": 2470
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2622,
      "step": 2471
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3243,
      "step": 2472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.1913,
      "step": 2473
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3985,
      "step": 2474
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2725,
      "step": 2475
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2527,
      "step": 2476
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3044,
      "step": 2477
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3308,
      "step": 2478
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.2127,
      "step": 2479
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4088,
      "step": 2480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4473,
      "step": 2481
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3835,
      "step": 2482
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.444,
      "step": 2483
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.3235,
      "step": 2484
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2917,
      "step": 2485
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2218,
      "step": 2486
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.434,
      "step": 2487
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3013,
      "step": 2488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3588,
      "step": 2489
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2785,
      "step": 2490
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2613,
      "step": 2491
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3019,
      "step": 2492
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2838,
      "step": 2493
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2999,
      "step": 2494
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4015,
      "step": 2495
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2963,
      "step": 2496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2522,
      "step": 2497
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2661,
      "step": 2498
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2917,
      "step": 2499
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3246,
      "step": 2500
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2823,
      "step": 2501
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3439,
      "step": 2502
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3459,
      "step": 2503
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2899,
      "step": 2504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2466,
      "step": 2505
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2391,
      "step": 2506
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3467,
      "step": 2507
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3835,
      "step": 2508
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3069,
      "step": 2509
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3175,
      "step": 2510
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2589,
      "step": 2511
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3573,
      "step": 2512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2977,
      "step": 2513
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4061,
      "step": 2514
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3099,
      "step": 2515
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3567,
      "step": 2516
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2306,
      "step": 2517
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5187,
      "step": 2518
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.356,
      "step": 2519
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2303,
      "step": 2520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2318,
      "step": 2521
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2958,
      "step": 2522
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3053,
      "step": 2523
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3214,
      "step": 2524
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3287,
      "step": 2525
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.2647,
      "step": 2526
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.602,
      "step": 2527
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3911,
      "step": 2528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4484,
      "step": 2529
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3538,
      "step": 2530
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3536,
      "step": 2531
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3169,
      "step": 2532
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3748,
      "step": 2533
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3992,
      "step": 2534
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3609,
      "step": 2535
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3587,
      "step": 2536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3127,
      "step": 2537
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5788,
      "step": 2538
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3057,
      "step": 2539
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3649,
      "step": 2540
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3006,
      "step": 2541
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.345,
      "step": 2542
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3167,
      "step": 2543
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4445,
      "step": 2544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.698,
      "step": 2545
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4337,
      "step": 2546
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3001,
      "step": 2547
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3034,
      "step": 2548
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2952,
      "step": 2549
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3464,
      "step": 2550
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2288,
      "step": 2551
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4369,
      "step": 2552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2054,
      "step": 2553
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4593,
      "step": 2554
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2664,
      "step": 2555
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2814,
      "step": 2556
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3875,
      "step": 2557
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5041,
      "step": 2558
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4504,
      "step": 2559
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2967,
      "step": 2560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5376,
      "step": 2561
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.315,
      "step": 2562
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4601,
      "step": 2563
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2925,
      "step": 2564
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4096,
      "step": 2565
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3639,
      "step": 2566
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3403,
      "step": 2567
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4106,
      "step": 2568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3305,
      "step": 2569
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2637,
      "step": 2570
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3166,
      "step": 2571
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3455,
      "step": 2572
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2861,
      "step": 2573
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2282,
      "step": 2574
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3501,
      "step": 2575
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3426,
      "step": 2576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2824,
      "step": 2577
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2988,
      "step": 2578
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3268,
      "step": 2579
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3739,
      "step": 2580
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.2955,
      "step": 2581
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3213,
      "step": 2582
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2529,
      "step": 2583
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2927,
      "step": 2584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2911,
      "step": 2585
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.1988,
      "step": 2586
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2285,
      "step": 2587
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5163,
      "step": 2588
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3237,
      "step": 2589
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3787,
      "step": 2590
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5032,
      "step": 2591
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2892,
      "step": 2592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3186,
      "step": 2593
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3112,
      "step": 2594
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2931,
      "step": 2595
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.289,
      "step": 2596
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.341,
      "step": 2597
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2766,
      "step": 2598
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3678,
      "step": 2599
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2661,
      "step": 2600
    },
    {
      "epoch": 0.54,
      "eval_loss": 0.32317888736724854,
      "eval_runtime": 159.3812,
      "eval_samples_per_second": 6.274,
      "eval_steps_per_second": 3.137,
      "step": 2600
    },
    {
      "epoch": 0.54,
      "mmlu_eval_accuracy": 0.5886664721106151,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "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": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.5454545454545454,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5769230769230769,
      "mmlu_eval_accuracy_econometrics": 0.4166666666666667,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3902439024390244,
      "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": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5116279069767442,
      "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": 0.8166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.5769230769230769,
      "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": 0.7272727272727273,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7674418604651163,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.28,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.6129032258064516,
      "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": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.3956259231109844,
      "step": 2600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2678,
      "step": 2601
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4585,
      "step": 2602
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3735,
      "step": 2603
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2919,
      "step": 2604
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4027,
      "step": 2605
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3543,
      "step": 2606
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3777,
      "step": 2607
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3948,
      "step": 2608
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3545,
      "step": 2609
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3116,
      "step": 2610
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.1999,
      "step": 2611
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4201,
      "step": 2612
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2551,
      "step": 2613
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2756,
      "step": 2614
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.456,
      "step": 2615
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2976,
      "step": 2616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2379,
      "step": 2617
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4378,
      "step": 2618
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3724,
      "step": 2619
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2857,
      "step": 2620
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3452,
      "step": 2621
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3682,
      "step": 2622
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3141,
      "step": 2623
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2875,
      "step": 2624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3781,
      "step": 2625
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2636,
      "step": 2626
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2781,
      "step": 2627
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2638,
      "step": 2628
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.2784,
      "step": 2629
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3024,
      "step": 2630
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4735,
      "step": 2631
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.2561,
      "step": 2632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5649,
      "step": 2633
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3802,
      "step": 2634
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4946,
      "step": 2635
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.2683,
      "step": 2636
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.2454,
      "step": 2637
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3362,
      "step": 2638
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3168,
      "step": 2639
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3256,
      "step": 2640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.2517,
      "step": 2641
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.2664,
      "step": 2642
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3457,
      "step": 2643
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3412,
      "step": 2644
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.2342,
      "step": 2645
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4006,
      "step": 2646
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4052,
      "step": 2647
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3916,
      "step": 2648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3251,
      "step": 2649
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.2646,
      "step": 2650
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3031,
      "step": 2651
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3778,
      "step": 2652
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.314,
      "step": 2653
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.2823,
      "step": 2654
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3627,
      "step": 2655
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3056,
      "step": 2656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.378,
      "step": 2657
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3023,
      "step": 2658
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3121,
      "step": 2659
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3162,
      "step": 2660
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3201,
      "step": 2661
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.2743,
      "step": 2662
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3499,
      "step": 2663
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3061,
      "step": 2664
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.2824,
      "step": 2665
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3073,
      "step": 2666
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.2879,
      "step": 2667
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3907,
      "step": 2668
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3019,
      "step": 2669
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.2857,
      "step": 2670
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3516,
      "step": 2671
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3112,
      "step": 2672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3155,
      "step": 2673
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3334,
      "step": 2674
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4611,
      "step": 2675
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3095,
      "step": 2676
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3655,
      "step": 2677
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3157,
      "step": 2678
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2491,
      "step": 2679
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3165,
      "step": 2680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2449,
      "step": 2681
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3033,
      "step": 2682
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4412,
      "step": 2683
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3236,
      "step": 2684
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.331,
      "step": 2685
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2731,
      "step": 2686
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2812,
      "step": 2687
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3318,
      "step": 2688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3496,
      "step": 2689
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3983,
      "step": 2690
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4462,
      "step": 2691
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3031,
      "step": 2692
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3074,
      "step": 2693
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3059,
      "step": 2694
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3372,
      "step": 2695
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3593,
      "step": 2696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2743,
      "step": 2697
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2557,
      "step": 2698
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4226,
      "step": 2699
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2603,
      "step": 2700
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.334,
      "step": 2701
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3246,
      "step": 2702
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2822,
      "step": 2703
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3369,
      "step": 2704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3663,
      "step": 2705
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4713,
      "step": 2706
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2944,
      "step": 2707
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.1726,
      "step": 2708
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6301,
      "step": 2709
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2985,
      "step": 2710
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3106,
      "step": 2711
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2489,
      "step": 2712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2474,
      "step": 2713
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.511,
      "step": 2714
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2863,
      "step": 2715
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4146,
      "step": 2716
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2697,
      "step": 2717
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2762,
      "step": 2718
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2343,
      "step": 2719
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2943,
      "step": 2720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2884,
      "step": 2721
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2578,
      "step": 2722
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.3623,
      "step": 2723
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2745,
      "step": 2724
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.2399,
      "step": 2725
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.305,
      "step": 2726
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3361,
      "step": 2727
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2688,
      "step": 2728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3104,
      "step": 2729
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3407,
      "step": 2730
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3809,
      "step": 2731
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2905,
      "step": 2732
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2296,
      "step": 2733
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3449,
      "step": 2734
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3099,
      "step": 2735
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3178,
      "step": 2736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3028,
      "step": 2737
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3041,
      "step": 2738
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2653,
      "step": 2739
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4068,
      "step": 2740
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3737,
      "step": 2741
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2743,
      "step": 2742
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3368,
      "step": 2743
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3911,
      "step": 2744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4049,
      "step": 2745
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4572,
      "step": 2746
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3239,
      "step": 2747
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2193,
      "step": 2748
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2697,
      "step": 2749
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2059,
      "step": 2750
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2617,
      "step": 2751
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.367,
      "step": 2752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3015,
      "step": 2753
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3626,
      "step": 2754
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2611,
      "step": 2755
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2673,
      "step": 2756
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3004,
      "step": 2757
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3072,
      "step": 2758
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.1914,
      "step": 2759
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2531,
      "step": 2760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2957,
      "step": 2761
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2423,
      "step": 2762
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2032,
      "step": 2763
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3063,
      "step": 2764
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3992,
      "step": 2765
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4091,
      "step": 2766
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2875,
      "step": 2767
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3014,
      "step": 2768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3102,
      "step": 2769
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.2961,
      "step": 2770
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3288,
      "step": 2771
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3461,
      "step": 2772
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5831,
      "step": 2773
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4181,
      "step": 2774
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2652,
      "step": 2775
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4437,
      "step": 2776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3924,
      "step": 2777
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2972,
      "step": 2778
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3092,
      "step": 2779
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4792,
      "step": 2780
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2436,
      "step": 2781
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3512,
      "step": 2782
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2933,
      "step": 2783
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2718,
      "step": 2784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2778,
      "step": 2785
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3124,
      "step": 2786
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4131,
      "step": 2787
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2983,
      "step": 2788
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2769,
      "step": 2789
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2323,
      "step": 2790
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2819,
      "step": 2791
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3174,
      "step": 2792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2047,
      "step": 2793
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3036,
      "step": 2794
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4808,
      "step": 2795
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2457,
      "step": 2796
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2952,
      "step": 2797
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2862,
      "step": 2798
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3317,
      "step": 2799
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2909,
      "step": 2800
    },
    {
      "epoch": 0.58,
      "eval_loss": 0.3221041262149811,
      "eval_runtime": 159.3141,
      "eval_samples_per_second": 6.277,
      "eval_steps_per_second": 3.138,
      "step": 2800
    },
    {
      "epoch": 0.58,
      "mmlu_eval_accuracy": 0.5887883659733487,
      "mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "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": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.4166666666666667,
      "mmlu_eval_accuracy_electrical_engineering": 0.6875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
      "mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.5625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.85,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "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": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.32,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.7647058823529411,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_law": 0.38235294117647056,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "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": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.6111111111111112,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.6431818206845625,
      "step": 2800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2481,
      "step": 2801
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3151,
      "step": 2802
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2285,
      "step": 2803
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3238,
      "step": 2804
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3991,
      "step": 2805
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3854,
      "step": 2806
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2339,
      "step": 2807
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3471,
      "step": 2808
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5219,
      "step": 2809
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.359,
      "step": 2810
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2925,
      "step": 2811
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2293,
      "step": 2812
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3684,
      "step": 2813
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4446,
      "step": 2814
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3156,
      "step": 2815
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2646,
      "step": 2816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4515,
      "step": 2817
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3486,
      "step": 2818
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3476,
      "step": 2819
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2304,
      "step": 2820
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.2494,
      "step": 2821
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3406,
      "step": 2822
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3075,
      "step": 2823
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2964,
      "step": 2824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4507,
      "step": 2825
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3524,
      "step": 2826
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2175,
      "step": 2827
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3832,
      "step": 2828
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2061,
      "step": 2829
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3479,
      "step": 2830
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2745,
      "step": 2831
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5658,
      "step": 2832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4696,
      "step": 2833
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3608,
      "step": 2834
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3105,
      "step": 2835
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2368,
      "step": 2836
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2259,
      "step": 2837
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2792,
      "step": 2838
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3624,
      "step": 2839
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5558,
      "step": 2840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2741,
      "step": 2841
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2669,
      "step": 2842
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3937,
      "step": 2843
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3666,
      "step": 2844
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3353,
      "step": 2845
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.247,
      "step": 2846
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5287,
      "step": 2847
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3173,
      "step": 2848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2324,
      "step": 2849
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3604,
      "step": 2850
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2669,
      "step": 2851
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4031,
      "step": 2852
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2566,
      "step": 2853
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3718,
      "step": 2854
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3168,
      "step": 2855
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2938,
      "step": 2856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2518,
      "step": 2857
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4042,
      "step": 2858
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2939,
      "step": 2859
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3135,
      "step": 2860
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.391,
      "step": 2861
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.322,
      "step": 2862
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2808,
      "step": 2863
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2463,
      "step": 2864
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4351,
      "step": 2865
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2931,
      "step": 2866
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2443,
      "step": 2867
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2537,
      "step": 2868
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.3095,
      "step": 2869
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.2343,
      "step": 2870
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2675,
      "step": 2871
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3368,
      "step": 2872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2672,
      "step": 2873
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4455,
      "step": 2874
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2955,
      "step": 2875
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.215,
      "step": 2876
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2089,
      "step": 2877
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2365,
      "step": 2878
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2871,
      "step": 2879
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3631,
      "step": 2880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3439,
      "step": 2881
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3307,
      "step": 2882
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4404,
      "step": 2883
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4213,
      "step": 2884
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3838,
      "step": 2885
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3515,
      "step": 2886
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3022,
      "step": 2887
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4811,
      "step": 2888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2369,
      "step": 2889
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3049,
      "step": 2890
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2903,
      "step": 2891
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4699,
      "step": 2892
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3309,
      "step": 2893
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2766,
      "step": 2894
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4932,
      "step": 2895
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3004,
      "step": 2896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2487,
      "step": 2897
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.212,
      "step": 2898
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.28,
      "step": 2899
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3032,
      "step": 2900
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2942,
      "step": 2901
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2782,
      "step": 2902
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2568,
      "step": 2903
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3819,
      "step": 2904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5146,
      "step": 2905
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3725,
      "step": 2906
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2537,
      "step": 2907
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3267,
      "step": 2908
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3363,
      "step": 2909
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3484,
      "step": 2910
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2962,
      "step": 2911
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3585,
      "step": 2912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3544,
      "step": 2913
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5711,
      "step": 2914
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3506,
      "step": 2915
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3223,
      "step": 2916
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.3151,
      "step": 2917
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.2864,
      "step": 2918
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2119,
      "step": 2919
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2347,
      "step": 2920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2935,
      "step": 2921
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3107,
      "step": 2922
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3292,
      "step": 2923
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3599,
      "step": 2924
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2804,
      "step": 2925
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3022,
      "step": 2926
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3786,
      "step": 2927
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3159,
      "step": 2928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4342,
      "step": 2929
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2707,
      "step": 2930
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3571,
      "step": 2931
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2523,
      "step": 2932
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3362,
      "step": 2933
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2107,
      "step": 2934
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3142,
      "step": 2935
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.1913,
      "step": 2936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2722,
      "step": 2937
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2782,
      "step": 2938
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2515,
      "step": 2939
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3216,
      "step": 2940
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3229,
      "step": 2941
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4138,
      "step": 2942
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3098,
      "step": 2943
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5058,
      "step": 2944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4088,
      "step": 2945
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2446,
      "step": 2946
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3867,
      "step": 2947
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2736,
      "step": 2948
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2272,
      "step": 2949
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2759,
      "step": 2950
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3634,
      "step": 2951
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3946,
      "step": 2952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3074,
      "step": 2953
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2773,
      "step": 2954
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2697,
      "step": 2955
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5257,
      "step": 2956
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3308,
      "step": 2957
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2752,
      "step": 2958
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3669,
      "step": 2959
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4004,
      "step": 2960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3442,
      "step": 2961
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3053,
      "step": 2962
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2421,
      "step": 2963
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3055,
      "step": 2964
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.2321,
      "step": 2965
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.3146,
      "step": 2966
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.1797,
      "step": 2967
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2522,
      "step": 2968
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3597,
      "step": 2969
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3146,
      "step": 2970
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4288,
      "step": 2971
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.261,
      "step": 2972
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5056,
      "step": 2973
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2448,
      "step": 2974
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.308,
      "step": 2975
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2809,
      "step": 2976
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3925,
      "step": 2977
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2372,
      "step": 2978
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2895,
      "step": 2979
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5911,
      "step": 2980
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4727,
      "step": 2981
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2818,
      "step": 2982
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.1952,
      "step": 2983
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3198,
      "step": 2984
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3814,
      "step": 2985
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3265,
      "step": 2986
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3845,
      "step": 2987
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2867,
      "step": 2988
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2801,
      "step": 2989
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.285,
      "step": 2990
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4482,
      "step": 2991
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2426,
      "step": 2992
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3867,
      "step": 2993
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2437,
      "step": 2994
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3322,
      "step": 2995
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2477,
      "step": 2996
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5995,
      "step": 2997
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2331,
      "step": 2998
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3155,
      "step": 2999
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3,
      "step": 3000
    },
    {
      "epoch": 0.62,
      "eval_loss": 0.3218897581100464,
      "eval_runtime": 159.4209,
      "eval_samples_per_second": 6.273,
      "eval_steps_per_second": 3.136,
      "step": 3000
    },
    {
      "epoch": 0.62,
      "mmlu_eval_accuracy": 0.5784152125467259,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "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": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.6153846153846154,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "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": 0.22727272727272727,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.6511627906976745,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6923076923076923,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.8,
      "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": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 0.8181818181818182,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.36,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.7714285714285715,
      "mmlu_eval_accuracy_professional_accounting": 0.3548387096774194,
      "mmlu_eval_accuracy_professional_law": 0.3588235294117647,
      "mmlu_eval_accuracy_professional_medicine": 0.5806451612903226,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.2072728029586006,
      "step": 3000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2474,
      "step": 3001
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3922,
      "step": 3002
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3544,
      "step": 3003
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2501,
      "step": 3004
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3207,
      "step": 3005
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3253,
      "step": 3006
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2648,
      "step": 3007
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2783,
      "step": 3008
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4098,
      "step": 3009
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.2873,
      "step": 3010
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3339,
      "step": 3011
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3701,
      "step": 3012
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3005,
      "step": 3013
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.286,
      "step": 3014
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3578,
      "step": 3015
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2828,
      "step": 3016
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2962,
      "step": 3017
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2524,
      "step": 3018
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3172,
      "step": 3019
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.202,
      "step": 3020
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3519,
      "step": 3021
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3748,
      "step": 3022
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3567,
      "step": 3023
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3728,
      "step": 3024
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3506,
      "step": 3025
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3118,
      "step": 3026
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2795,
      "step": 3027
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2844,
      "step": 3028
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5738,
      "step": 3029
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3751,
      "step": 3030
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3197,
      "step": 3031
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2727,
      "step": 3032
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2789,
      "step": 3033
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4277,
      "step": 3034
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3234,
      "step": 3035
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2465,
      "step": 3036
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3172,
      "step": 3037
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3603,
      "step": 3038
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2619,
      "step": 3039
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3494,
      "step": 3040
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2588,
      "step": 3041
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.305,
      "step": 3042
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3623,
      "step": 3043
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2019,
      "step": 3044
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2957,
      "step": 3045
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2951,
      "step": 3046
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3815,
      "step": 3047
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.334,
      "step": 3048
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3307,
      "step": 3049
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2318,
      "step": 3050
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2778,
      "step": 3051
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3514,
      "step": 3052
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3774,
      "step": 3053
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5391,
      "step": 3054
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.255,
      "step": 3055
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.2266,
      "step": 3056
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4524,
      "step": 3057
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4524,
      "step": 3058
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4258,
      "step": 3059
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.257,
      "step": 3060
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.21,
      "step": 3061
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3487,
      "step": 3062
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3512,
      "step": 3063
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3409,
      "step": 3064
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2859,
      "step": 3065
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4384,
      "step": 3066
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2503,
      "step": 3067
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3583,
      "step": 3068
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2334,
      "step": 3069
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.237,
      "step": 3070
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2435,
      "step": 3071
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3096,
      "step": 3072
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2093,
      "step": 3073
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3294,
      "step": 3074
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5412,
      "step": 3075
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3689,
      "step": 3076
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2384,
      "step": 3077
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3083,
      "step": 3078
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4412,
      "step": 3079
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.291,
      "step": 3080
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3309,
      "step": 3081
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3104,
      "step": 3082
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2883,
      "step": 3083
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3862,
      "step": 3084
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2315,
      "step": 3085
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.31,
      "step": 3086
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2531,
      "step": 3087
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2349,
      "step": 3088
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3617,
      "step": 3089
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.275,
      "step": 3090
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2799,
      "step": 3091
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2617,
      "step": 3092
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3382,
      "step": 3093
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.312,
      "step": 3094
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3595,
      "step": 3095
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2423,
      "step": 3096
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3622,
      "step": 3097
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2392,
      "step": 3098
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3433,
      "step": 3099
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3191,
      "step": 3100
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.1908,
      "step": 3101
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3544,
      "step": 3102
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2705,
      "step": 3103
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2841,
      "step": 3104
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2961,
      "step": 3105
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3494,
      "step": 3106
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2679,
      "step": 3107
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2973,
      "step": 3108
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3135,
      "step": 3109
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3588,
      "step": 3110
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.2548,
      "step": 3111
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3941,
      "step": 3112
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3031,
      "step": 3113
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3867,
      "step": 3114
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2433,
      "step": 3115
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3268,
      "step": 3116
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2819,
      "step": 3117
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2888,
      "step": 3118
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3869,
      "step": 3119
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2939,
      "step": 3120
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3754,
      "step": 3121
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2376,
      "step": 3122
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3029,
      "step": 3123
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3653,
      "step": 3124
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3419,
      "step": 3125
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4735,
      "step": 3126
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3326,
      "step": 3127
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3497,
      "step": 3128
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3756,
      "step": 3129
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2858,
      "step": 3130
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3627,
      "step": 3131
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2844,
      "step": 3132
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2873,
      "step": 3133
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2185,
      "step": 3134
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3579,
      "step": 3135
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3395,
      "step": 3136
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3161,
      "step": 3137
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3152,
      "step": 3138
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3861,
      "step": 3139
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3083,
      "step": 3140
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2551,
      "step": 3141
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4283,
      "step": 3142
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3133,
      "step": 3143
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2477,
      "step": 3144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2909,
      "step": 3145
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2906,
      "step": 3146
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2562,
      "step": 3147
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2515,
      "step": 3148
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.466,
      "step": 3149
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2595,
      "step": 3150
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3271,
      "step": 3151
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2985,
      "step": 3152
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3677,
      "step": 3153
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2957,
      "step": 3154
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3431,
      "step": 3155
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3116,
      "step": 3156
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.3411,
      "step": 3157
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2001,
      "step": 3158
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2663,
      "step": 3159
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.2966,
      "step": 3160
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3319,
      "step": 3161
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2592,
      "step": 3162
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2404,
      "step": 3163
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.269,
      "step": 3164
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2625,
      "step": 3165
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2651,
      "step": 3166
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3386,
      "step": 3167
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3527,
      "step": 3168
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2425,
      "step": 3169
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3075,
      "step": 3170
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3985,
      "step": 3171
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2475,
      "step": 3172
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2687,
      "step": 3173
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2271,
      "step": 3174
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2175,
      "step": 3175
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2788,
      "step": 3176
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2859,
      "step": 3177
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2354,
      "step": 3178
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5822,
      "step": 3179
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3464,
      "step": 3180
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3518,
      "step": 3181
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3995,
      "step": 3182
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5163,
      "step": 3183
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2157,
      "step": 3184
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3671,
      "step": 3185
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4704,
      "step": 3186
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2982,
      "step": 3187
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3178,
      "step": 3188
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2658,
      "step": 3189
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3225,
      "step": 3190
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3462,
      "step": 3191
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.293,
      "step": 3192
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.37,
      "step": 3193
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3485,
      "step": 3194
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5336,
      "step": 3195
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.295,
      "step": 3196
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2935,
      "step": 3197
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4975,
      "step": 3198
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3271,
      "step": 3199
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2364,
      "step": 3200
    },
    {
      "epoch": 0.66,
      "eval_loss": 0.3187420666217804,
      "eval_runtime": 159.3571,
      "eval_samples_per_second": 6.275,
      "eval_steps_per_second": 3.138,
      "step": 3200
    },
    {
      "epoch": 0.66,
      "mmlu_eval_accuracy": 0.5846092789618116,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6551724137931034,
      "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": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.5769230769230769,
      "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": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.627906976744186,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "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": 0.43478260869565216,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 0.8181818181818182,
      "mmlu_eval_accuracy_miscellaneous": 0.7674418604651163,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.36,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.7714285714285715,
      "mmlu_eval_accuracy_professional_accounting": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_law": 0.38235294117647056,
      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_psychology": 0.5797101449275363,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.462505467877998,
      "step": 3200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2365,
      "step": 3201
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3383,
      "step": 3202
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2863,
      "step": 3203
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3211,
      "step": 3204
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3625,
      "step": 3205
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.2823,
      "step": 3206
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.3256,
      "step": 3207
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5096,
      "step": 3208
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2715,
      "step": 3209
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2162,
      "step": 3210
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2971,
      "step": 3211
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6382,
      "step": 3212
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3194,
      "step": 3213
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2965,
      "step": 3214
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.1775,
      "step": 3215
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3829,
      "step": 3216
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2208,
      "step": 3217
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2881,
      "step": 3218
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3746,
      "step": 3219
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3232,
      "step": 3220
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4368,
      "step": 3221
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.311,
      "step": 3222
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5344,
      "step": 3223
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2694,
      "step": 3224
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2215,
      "step": 3225
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3142,
      "step": 3226
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.298,
      "step": 3227
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2659,
      "step": 3228
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.242,
      "step": 3229
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3139,
      "step": 3230
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2806,
      "step": 3231
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2574,
      "step": 3232
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3475,
      "step": 3233
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3136,
      "step": 3234
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2252,
      "step": 3235
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.457,
      "step": 3236
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2857,
      "step": 3237
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2421,
      "step": 3238
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4159,
      "step": 3239
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2882,
      "step": 3240
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3636,
      "step": 3241
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2557,
      "step": 3242
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3719,
      "step": 3243
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3639,
      "step": 3244
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3911,
      "step": 3245
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.1792,
      "step": 3246
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3653,
      "step": 3247
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2253,
      "step": 3248
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4914,
      "step": 3249
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2519,
      "step": 3250
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3078,
      "step": 3251
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3235,
      "step": 3252
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2748,
      "step": 3253
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3662,
      "step": 3254
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.3839,
      "step": 3255
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.2989,
      "step": 3256
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3032,
      "step": 3257
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.194,
      "step": 3258
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3145,
      "step": 3259
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3577,
      "step": 3260
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2137,
      "step": 3261
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3938,
      "step": 3262
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.4112,
      "step": 3263
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3794,
      "step": 3264
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2803,
      "step": 3265
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2657,
      "step": 3266
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5806,
      "step": 3267
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2875,
      "step": 3268
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2885,
      "step": 3269
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2828,
      "step": 3270
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2935,
      "step": 3271
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.4984,
      "step": 3272
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.1892,
      "step": 3273
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2136,
      "step": 3274
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3323,
      "step": 3275
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3231,
      "step": 3276
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2762,
      "step": 3277
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3424,
      "step": 3278
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3046,
      "step": 3279
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3986,
      "step": 3280
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2907,
      "step": 3281
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3279,
      "step": 3282
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2459,
      "step": 3283
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.339,
      "step": 3284
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.1838,
      "step": 3285
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3094,
      "step": 3286
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2783,
      "step": 3287
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.4122,
      "step": 3288
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2623,
      "step": 3289
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3218,
      "step": 3290
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.1788,
      "step": 3291
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2169,
      "step": 3292
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3265,
      "step": 3293
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.4319,
      "step": 3294
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.4227,
      "step": 3295
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2778,
      "step": 3296
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3112,
      "step": 3297
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2991,
      "step": 3298
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.2131,
      "step": 3299
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.264,
      "step": 3300
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3058,
      "step": 3301
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3031,
      "step": 3302
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3276,
      "step": 3303
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.313,
      "step": 3304
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3496,
      "step": 3305
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2689,
      "step": 3306
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2402,
      "step": 3307
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2697,
      "step": 3308
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3915,
      "step": 3309
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2519,
      "step": 3310
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3584,
      "step": 3311
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3734,
      "step": 3312
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2629,
      "step": 3313
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.307,
      "step": 3314
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.465,
      "step": 3315
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5002,
      "step": 3316
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2761,
      "step": 3317
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2337,
      "step": 3318
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2424,
      "step": 3319
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4512,
      "step": 3320
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.1996,
      "step": 3321
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4767,
      "step": 3322
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3616,
      "step": 3323
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3384,
      "step": 3324
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2649,
      "step": 3325
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3358,
      "step": 3326
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2428,
      "step": 3327
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4393,
      "step": 3328
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2343,
      "step": 3329
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3698,
      "step": 3330
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3455,
      "step": 3331
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2294,
      "step": 3332
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4358,
      "step": 3333
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3228,
      "step": 3334
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3534,
      "step": 3335
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3445,
      "step": 3336
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4302,
      "step": 3337
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3148,
      "step": 3338
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2754,
      "step": 3339
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2926,
      "step": 3340
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.433,
      "step": 3341
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3184,
      "step": 3342
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3343,
      "step": 3343
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2634,
      "step": 3344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2762,
      "step": 3345
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2829,
      "step": 3346
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2739,
      "step": 3347
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2829,
      "step": 3348
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3758,
      "step": 3349
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2496,
      "step": 3350
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3308,
      "step": 3351
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.2526,
      "step": 3352
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.3582,
      "step": 3353
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2688,
      "step": 3354
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3031,
      "step": 3355
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2956,
      "step": 3356
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3148,
      "step": 3357
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3955,
      "step": 3358
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4213,
      "step": 3359
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2681,
      "step": 3360
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.247,
      "step": 3361
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3714,
      "step": 3362
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3271,
      "step": 3363
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4888,
      "step": 3364
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5578,
      "step": 3365
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3036,
      "step": 3366
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3428,
      "step": 3367
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.368,
      "step": 3368
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3719,
      "step": 3369
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2533,
      "step": 3370
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.283,
      "step": 3371
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.1922,
      "step": 3372
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3214,
      "step": 3373
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.293,
      "step": 3374
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2407,
      "step": 3375
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3634,
      "step": 3376
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2536,
      "step": 3377
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3672,
      "step": 3378
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4147,
      "step": 3379
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3176,
      "step": 3380
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3054,
      "step": 3381
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2366,
      "step": 3382
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2597,
      "step": 3383
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2378,
      "step": 3384
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4275,
      "step": 3385
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3099,
      "step": 3386
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2747,
      "step": 3387
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3998,
      "step": 3388
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3474,
      "step": 3389
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.292,
      "step": 3390
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2809,
      "step": 3391
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2459,
      "step": 3392
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2906,
      "step": 3393
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3447,
      "step": 3394
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.2906,
      "step": 3395
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3288,
      "step": 3396
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4326,
      "step": 3397
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.1976,
      "step": 3398
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4197,
      "step": 3399
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.242,
      "step": 3400
    },
    {
      "epoch": 0.7,
      "eval_loss": 0.317775160074234,
      "eval_runtime": 159.2501,
      "eval_samples_per_second": 6.279,
      "eval_steps_per_second": 3.14,
      "step": 3400
    },
    {
      "epoch": 0.7,
      "mmlu_eval_accuracy": 0.5788751603061874,
      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.6875,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "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": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.5769230769230769,
      "mmlu_eval_accuracy_econometrics": 0.3333333333333333,
      "mmlu_eval_accuracy_electrical_engineering": 0.4375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.43902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.07142857142857142,
      "mmlu_eval_accuracy_global_facts": 0.6,
      "mmlu_eval_accuracy_high_school_biology": 0.65625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.627906976744186,
      "mmlu_eval_accuracy_high_school_mathematics": 0.13793103448275862,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6923076923076923,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "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": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "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": 0.42,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.6764705882352942,
      "mmlu_eval_accuracy_prehistory": 0.6285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_law": 0.38235294117647056,
      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_psychology": 0.5797101449275363,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.8636363636363636,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.5392296446132285,
      "step": 3400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3055,
      "step": 3401
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2808,
      "step": 3402
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5172,
      "step": 3403
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2518,
      "step": 3404
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3222,
      "step": 3405
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2141,
      "step": 3406
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2634,
      "step": 3407
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3122,
      "step": 3408
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5583,
      "step": 3409
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2961,
      "step": 3410
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2372,
      "step": 3411
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2844,
      "step": 3412
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3348,
      "step": 3413
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3394,
      "step": 3414
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2995,
      "step": 3415
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4542,
      "step": 3416
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3292,
      "step": 3417
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2841,
      "step": 3418
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.408,
      "step": 3419
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2691,
      "step": 3420
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2369,
      "step": 3421
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2773,
      "step": 3422
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.1933,
      "step": 3423
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4611,
      "step": 3424
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.265,
      "step": 3425
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4634,
      "step": 3426
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3778,
      "step": 3427
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2698,
      "step": 3428
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4339,
      "step": 3429
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2442,
      "step": 3430
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2204,
      "step": 3431
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.1925,
      "step": 3432
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2301,
      "step": 3433
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2918,
      "step": 3434
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3765,
      "step": 3435
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3418,
      "step": 3436
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.401,
      "step": 3437
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2796,
      "step": 3438
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2919,
      "step": 3439
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2813,
      "step": 3440
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.32,
      "step": 3441
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3251,
      "step": 3442
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3282,
      "step": 3443
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3083,
      "step": 3444
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3754,
      "step": 3445
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.3666,
      "step": 3446
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.217,
      "step": 3447
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2809,
      "step": 3448
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.2752,
      "step": 3449
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.3091,
      "step": 3450
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2124,
      "step": 3451
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2657,
      "step": 3452
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2922,
      "step": 3453
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4305,
      "step": 3454
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2642,
      "step": 3455
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.3324,
      "step": 3456
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.411,
      "step": 3457
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2891,
      "step": 3458
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2974,
      "step": 3459
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2711,
      "step": 3460
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2677,
      "step": 3461
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2387,
      "step": 3462
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.3035,
      "step": 3463
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2287,
      "step": 3464
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2115,
      "step": 3465
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2512,
      "step": 3466
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2953,
      "step": 3467
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.28,
      "step": 3468
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.3238,
      "step": 3469
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2282,
      "step": 3470
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2512,
      "step": 3471
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2786,
      "step": 3472
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2247,
      "step": 3473
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2612,
      "step": 3474
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4461,
      "step": 3475
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2153,
      "step": 3476
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2736,
      "step": 3477
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2708,
      "step": 3478
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6198,
      "step": 3479
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2457,
      "step": 3480
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.3431,
      "step": 3481
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5975,
      "step": 3482
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2252,
      "step": 3483
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.309,
      "step": 3484
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.248,
      "step": 3485
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2653,
      "step": 3486
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.342,
      "step": 3487
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2068,
      "step": 3488
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.3483,
      "step": 3489
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2928,
      "step": 3490
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.467,
      "step": 3491
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2291,
      "step": 3492
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2862,
      "step": 3493
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4046,
      "step": 3494
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2116,
      "step": 3495
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2594,
      "step": 3496
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.2865,
      "step": 3497
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3751,
      "step": 3498
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2988,
      "step": 3499
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2142,
      "step": 3500
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2802,
      "step": 3501
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2757,
      "step": 3502
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3317,
      "step": 3503
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2867,
      "step": 3504
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2555,
      "step": 3505
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.213,
      "step": 3506
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2267,
      "step": 3507
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2681,
      "step": 3508
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2767,
      "step": 3509
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4298,
      "step": 3510
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3203,
      "step": 3511
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.334,
      "step": 3512
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3449,
      "step": 3513
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2899,
      "step": 3514
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3357,
      "step": 3515
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3236,
      "step": 3516
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2881,
      "step": 3517
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2813,
      "step": 3518
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2708,
      "step": 3519
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3969,
      "step": 3520
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5111,
      "step": 3521
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3021,
      "step": 3522
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3154,
      "step": 3523
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2301,
      "step": 3524
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2393,
      "step": 3525
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2602,
      "step": 3526
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3499,
      "step": 3527
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3376,
      "step": 3528
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2692,
      "step": 3529
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2624,
      "step": 3530
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2827,
      "step": 3531
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2663,
      "step": 3532
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3301,
      "step": 3533
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3347,
      "step": 3534
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.283,
      "step": 3535
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3495,
      "step": 3536
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2422,
      "step": 3537
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2795,
      "step": 3538
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2115,
      "step": 3539
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2606,
      "step": 3540
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2503,
      "step": 3541
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3016,
      "step": 3542
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2469,
      "step": 3543
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3927,
      "step": 3544
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.2755,
      "step": 3545
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4764,
      "step": 3546
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.248,
      "step": 3547
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2405,
      "step": 3548
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.1705,
      "step": 3549
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2926,
      "step": 3550
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4283,
      "step": 3551
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2963,
      "step": 3552
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.266,
      "step": 3553
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2044,
      "step": 3554
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2264,
      "step": 3555
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4746,
      "step": 3556
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2545,
      "step": 3557
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2377,
      "step": 3558
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.1879,
      "step": 3559
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.288,
      "step": 3560
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2132,
      "step": 3561
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3289,
      "step": 3562
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.264,
      "step": 3563
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2315,
      "step": 3564
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3023,
      "step": 3565
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3002,
      "step": 3566
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2531,
      "step": 3567
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4957,
      "step": 3568
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3545,
      "step": 3569
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3212,
      "step": 3570
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3107,
      "step": 3571
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3624,
      "step": 3572
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.226,
      "step": 3573
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3043,
      "step": 3574
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3459,
      "step": 3575
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2944,
      "step": 3576
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2459,
      "step": 3577
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2376,
      "step": 3578
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3575,
      "step": 3579
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.317,
      "step": 3580
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5285,
      "step": 3581
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4566,
      "step": 3582
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3092,
      "step": 3583
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3041,
      "step": 3584
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2704,
      "step": 3585
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3363,
      "step": 3586
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3262,
      "step": 3587
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3094,
      "step": 3588
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3011,
      "step": 3589
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3291,
      "step": 3590
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3309,
      "step": 3591
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.3383,
      "step": 3592
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.324,
      "step": 3593
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.2004,
      "step": 3594
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2694,
      "step": 3595
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3555,
      "step": 3596
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3389,
      "step": 3597
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.275,
      "step": 3598
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2717,
      "step": 3599
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2987,
      "step": 3600
    },
    {
      "epoch": 0.75,
      "eval_loss": 0.316895455121994,
      "eval_runtime": 159.3839,
      "eval_samples_per_second": 6.274,
      "eval_steps_per_second": 3.137,
      "step": 3600
    },
    {
      "epoch": 0.75,
      "mmlu_eval_accuracy": 0.5955420579063287,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6551724137931034,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "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": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.6153846153846154,
      "mmlu_eval_accuracy_econometrics": 0.3333333333333333,
      "mmlu_eval_accuracy_electrical_engineering": 0.5,
      "mmlu_eval_accuracy_elementary_mathematics": 0.5121951219512195,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.6,
      "mmlu_eval_accuracy_high_school_biology": 0.59375,
      "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": 0.8181818181818182,
      "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": 0.5769230769230769,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "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": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7674418604651163,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.26,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.6857142857142857,
      "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": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.8636363636363636,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.4479418122659464,
      "step": 3600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3405,
      "step": 3601
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3697,
      "step": 3602
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2312,
      "step": 3603
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.1901,
      "step": 3604
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2932,
      "step": 3605
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2091,
      "step": 3606
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2126,
      "step": 3607
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4171,
      "step": 3608
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2984,
      "step": 3609
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2478,
      "step": 3610
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4204,
      "step": 3611
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2727,
      "step": 3612
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2377,
      "step": 3613
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3131,
      "step": 3614
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3481,
      "step": 3615
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3382,
      "step": 3616
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3299,
      "step": 3617
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2305,
      "step": 3618
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5367,
      "step": 3619
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2494,
      "step": 3620
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.276,
      "step": 3621
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.302,
      "step": 3622
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2883,
      "step": 3623
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2221,
      "step": 3624
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2018,
      "step": 3625
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2279,
      "step": 3626
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.419,
      "step": 3627
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2027,
      "step": 3628
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3623,
      "step": 3629
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2799,
      "step": 3630
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3456,
      "step": 3631
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4178,
      "step": 3632
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3203,
      "step": 3633
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2667,
      "step": 3634
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2483,
      "step": 3635
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3209,
      "step": 3636
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3154,
      "step": 3637
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2852,
      "step": 3638
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2075,
      "step": 3639
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4342,
      "step": 3640
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.2227,
      "step": 3641
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.273,
      "step": 3642
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3095,
      "step": 3643
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2319,
      "step": 3644
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5564,
      "step": 3645
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2146,
      "step": 3646
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2289,
      "step": 3647
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2843,
      "step": 3648
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2896,
      "step": 3649
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2367,
      "step": 3650
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2697,
      "step": 3651
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2682,
      "step": 3652
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3346,
      "step": 3653
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3099,
      "step": 3654
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.422,
      "step": 3655
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3287,
      "step": 3656
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2427,
      "step": 3657
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3068,
      "step": 3658
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4016,
      "step": 3659
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4019,
      "step": 3660
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2846,
      "step": 3661
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3447,
      "step": 3662
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2737,
      "step": 3663
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2379,
      "step": 3664
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3579,
      "step": 3665
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2856,
      "step": 3666
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2914,
      "step": 3667
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2759,
      "step": 3668
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3786,
      "step": 3669
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4391,
      "step": 3670
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.1435,
      "step": 3671
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3069,
      "step": 3672
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2337,
      "step": 3673
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2808,
      "step": 3674
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2561,
      "step": 3675
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.466,
      "step": 3676
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3777,
      "step": 3677
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3382,
      "step": 3678
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3249,
      "step": 3679
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2525,
      "step": 3680
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2219,
      "step": 3681
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2367,
      "step": 3682
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2601,
      "step": 3683
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3324,
      "step": 3684
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.25,
      "step": 3685
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3351,
      "step": 3686
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2843,
      "step": 3687
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3451,
      "step": 3688
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3062,
      "step": 3689
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.2916,
      "step": 3690
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2395,
      "step": 3691
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3078,
      "step": 3692
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2057,
      "step": 3693
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2255,
      "step": 3694
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2412,
      "step": 3695
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.273,
      "step": 3696
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3092,
      "step": 3697
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3206,
      "step": 3698
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3086,
      "step": 3699
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2891,
      "step": 3700
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.283,
      "step": 3701
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3462,
      "step": 3702
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2714,
      "step": 3703
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2814,
      "step": 3704
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3075,
      "step": 3705
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2344,
      "step": 3706
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4674,
      "step": 3707
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2976,
      "step": 3708
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.389,
      "step": 3709
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.335,
      "step": 3710
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2531,
      "step": 3711
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.374,
      "step": 3712
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3246,
      "step": 3713
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2227,
      "step": 3714
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3041,
      "step": 3715
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.296,
      "step": 3716
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3737,
      "step": 3717
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2299,
      "step": 3718
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.275,
      "step": 3719
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2908,
      "step": 3720
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3619,
      "step": 3721
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2925,
      "step": 3722
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3421,
      "step": 3723
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2749,
      "step": 3724
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.386,
      "step": 3725
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3349,
      "step": 3726
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3218,
      "step": 3727
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3807,
      "step": 3728
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.458,
      "step": 3729
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4571,
      "step": 3730
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2592,
      "step": 3731
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3578,
      "step": 3732
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2871,
      "step": 3733
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2289,
      "step": 3734
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.243,
      "step": 3735
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4938,
      "step": 3736
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3749,
      "step": 3737
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3383,
      "step": 3738
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.2373,
      "step": 3739
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3339,
      "step": 3740
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4161,
      "step": 3741
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3407,
      "step": 3742
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2818,
      "step": 3743
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.304,
      "step": 3744
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2828,
      "step": 3745
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3461,
      "step": 3746
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4329,
      "step": 3747
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3913,
      "step": 3748
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3264,
      "step": 3749
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3387,
      "step": 3750
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2464,
      "step": 3751
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3351,
      "step": 3752
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2421,
      "step": 3753
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2272,
      "step": 3754
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2902,
      "step": 3755
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2757,
      "step": 3756
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2773,
      "step": 3757
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.337,
      "step": 3758
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3045,
      "step": 3759
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2235,
      "step": 3760
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2584,
      "step": 3761
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4239,
      "step": 3762
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4423,
      "step": 3763
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2798,
      "step": 3764
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2385,
      "step": 3765
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2382,
      "step": 3766
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2156,
      "step": 3767
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.311,
      "step": 3768
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2642,
      "step": 3769
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3025,
      "step": 3770
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4322,
      "step": 3771
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2117,
      "step": 3772
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.352,
      "step": 3773
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2666,
      "step": 3774
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3625,
      "step": 3775
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2988,
      "step": 3776
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4504,
      "step": 3777
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.35,
      "step": 3778
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4691,
      "step": 3779
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.381,
      "step": 3780
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4083,
      "step": 3781
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2544,
      "step": 3782
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3152,
      "step": 3783
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4237,
      "step": 3784
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2567,
      "step": 3785
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3403,
      "step": 3786
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.2861,
      "step": 3787
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4286,
      "step": 3788
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2666,
      "step": 3789
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3084,
      "step": 3790
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2607,
      "step": 3791
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4458,
      "step": 3792
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5272,
      "step": 3793
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3737,
      "step": 3794
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3105,
      "step": 3795
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3025,
      "step": 3796
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3583,
      "step": 3797
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3232,
      "step": 3798
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2829,
      "step": 3799
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3597,
      "step": 3800
    },
    {
      "epoch": 0.79,
      "eval_loss": 0.31663838028907776,
      "eval_runtime": 159.3604,
      "eval_samples_per_second": 6.275,
      "eval_steps_per_second": 3.138,
      "step": 3800
    },
    {
      "epoch": 0.79,
      "mmlu_eval_accuracy": 0.5783087244203517,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
      "mmlu_eval_accuracy_college_medicine": 0.5454545454545454,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5769230769230769,
      "mmlu_eval_accuracy_econometrics": 0.3333333333333333,
      "mmlu_eval_accuracy_electrical_engineering": 0.4375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4878048780487805,
      "mmlu_eval_accuracy_formal_logic": 0.07142857142857142,
      "mmlu_eval_accuracy_global_facts": 0.6,
      "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": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5348837209302325,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "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": 0.4782608695652174,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "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": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.36,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.6285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_law": 0.37058823529411766,
      "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": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.5375341653590413,
      "step": 3800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3416,
      "step": 3801
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2586,
      "step": 3802
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.212,
      "step": 3803
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2522,
      "step": 3804
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2901,
      "step": 3805
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2171,
      "step": 3806
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2656,
      "step": 3807
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.316,
      "step": 3808
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5542,
      "step": 3809
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2535,
      "step": 3810
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.299,
      "step": 3811
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3042,
      "step": 3812
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2432,
      "step": 3813
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4792,
      "step": 3814
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2438,
      "step": 3815
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3403,
      "step": 3816
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2088,
      "step": 3817
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4703,
      "step": 3818
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3929,
      "step": 3819
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2533,
      "step": 3820
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2869,
      "step": 3821
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2747,
      "step": 3822
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3325,
      "step": 3823
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5278,
      "step": 3824
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2073,
      "step": 3825
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3516,
      "step": 3826
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3467,
      "step": 3827
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2729,
      "step": 3828
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.318,
      "step": 3829
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2979,
      "step": 3830
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.397,
      "step": 3831
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.3411,
      "step": 3832
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2984,
      "step": 3833
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2469,
      "step": 3834
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.2342,
      "step": 3835
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3959,
      "step": 3836
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3018,
      "step": 3837
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4221,
      "step": 3838
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3585,
      "step": 3839
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2136,
      "step": 3840
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3075,
      "step": 3841
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4248,
      "step": 3842
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2452,
      "step": 3843
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.247,
      "step": 3844
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2724,
      "step": 3845
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3005,
      "step": 3846
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2226,
      "step": 3847
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4856,
      "step": 3848
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2288,
      "step": 3849
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3401,
      "step": 3850
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3092,
      "step": 3851
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3361,
      "step": 3852
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3121,
      "step": 3853
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3513,
      "step": 3854
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3173,
      "step": 3855
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2741,
      "step": 3856
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2905,
      "step": 3857
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2296,
      "step": 3858
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2888,
      "step": 3859
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2066,
      "step": 3860
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2898,
      "step": 3861
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2695,
      "step": 3862
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5149,
      "step": 3863
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2794,
      "step": 3864
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.392,
      "step": 3865
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3778,
      "step": 3866
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2989,
      "step": 3867
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2369,
      "step": 3868
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3562,
      "step": 3869
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3492,
      "step": 3870
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3343,
      "step": 3871
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2408,
      "step": 3872
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3212,
      "step": 3873
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3431,
      "step": 3874
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3765,
      "step": 3875
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.1674,
      "step": 3876
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3737,
      "step": 3877
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2517,
      "step": 3878
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.384,
      "step": 3879
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3358,
      "step": 3880
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4137,
      "step": 3881
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2238,
      "step": 3882
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.2787,
      "step": 3883
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2555,
      "step": 3884
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.225,
      "step": 3885
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2987,
      "step": 3886
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2424,
      "step": 3887
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2594,
      "step": 3888
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2658,
      "step": 3889
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3273,
      "step": 3890
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2948,
      "step": 3891
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3367,
      "step": 3892
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2694,
      "step": 3893
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2044,
      "step": 3894
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.318,
      "step": 3895
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3533,
      "step": 3896
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4717,
      "step": 3897
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2615,
      "step": 3898
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3194,
      "step": 3899
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.442,
      "step": 3900
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2756,
      "step": 3901
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2847,
      "step": 3902
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3048,
      "step": 3903
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3006,
      "step": 3904
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.273,
      "step": 3905
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2329,
      "step": 3906
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2828,
      "step": 3907
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2364,
      "step": 3908
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4544,
      "step": 3909
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2347,
      "step": 3910
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2311,
      "step": 3911
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3004,
      "step": 3912
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3221,
      "step": 3913
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4362,
      "step": 3914
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.367,
      "step": 3915
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2067,
      "step": 3916
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3216,
      "step": 3917
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3709,
      "step": 3918
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.432,
      "step": 3919
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2913,
      "step": 3920
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4294,
      "step": 3921
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.1832,
      "step": 3922
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3076,
      "step": 3923
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2943,
      "step": 3924
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3004,
      "step": 3925
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2412,
      "step": 3926
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4029,
      "step": 3927
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2377,
      "step": 3928
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.2254,
      "step": 3929
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4391,
      "step": 3930
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3314,
      "step": 3931
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3348,
      "step": 3932
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2169,
      "step": 3933
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3602,
      "step": 3934
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.342,
      "step": 3935
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2627,
      "step": 3936
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.313,
      "step": 3937
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2905,
      "step": 3938
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5079,
      "step": 3939
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2494,
      "step": 3940
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2532,
      "step": 3941
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3855,
      "step": 3942
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2737,
      "step": 3943
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3014,
      "step": 3944
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3438,
      "step": 3945
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3003,
      "step": 3946
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3349,
      "step": 3947
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4988,
      "step": 3948
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2736,
      "step": 3949
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2689,
      "step": 3950
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3168,
      "step": 3951
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2619,
      "step": 3952
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3514,
      "step": 3953
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2214,
      "step": 3954
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.357,
      "step": 3955
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.28,
      "step": 3956
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4532,
      "step": 3957
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.322,
      "step": 3958
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3928,
      "step": 3959
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3189,
      "step": 3960
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3818,
      "step": 3961
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3436,
      "step": 3962
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2113,
      "step": 3963
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2033,
      "step": 3964
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2692,
      "step": 3965
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.1864,
      "step": 3966
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3974,
      "step": 3967
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2568,
      "step": 3968
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2341,
      "step": 3969
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.287,
      "step": 3970
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3006,
      "step": 3971
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3708,
      "step": 3972
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4491,
      "step": 3973
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2428,
      "step": 3974
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.2702,
      "step": 3975
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.265,
      "step": 3976
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.163,
      "step": 3977
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3015,
      "step": 3978
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.3299,
      "step": 3979
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.226,
      "step": 3980
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2953,
      "step": 3981
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2918,
      "step": 3982
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3287,
      "step": 3983
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3592,
      "step": 3984
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3319,
      "step": 3985
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2898,
      "step": 3986
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4053,
      "step": 3987
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3766,
      "step": 3988
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2908,
      "step": 3989
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3698,
      "step": 3990
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2264,
      "step": 3991
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2425,
      "step": 3992
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2725,
      "step": 3993
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3436,
      "step": 3994
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3453,
      "step": 3995
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2782,
      "step": 3996
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2684,
      "step": 3997
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2463,
      "step": 3998
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3929,
      "step": 3999
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3572,
      "step": 4000
    },
    {
      "epoch": 0.83,
      "eval_loss": 0.3149797022342682,
      "eval_runtime": 159.3285,
      "eval_samples_per_second": 6.276,
      "eval_steps_per_second": 3.138,
      "step": 4000
    },
    {
      "epoch": 0.83,
      "mmlu_eval_accuracy": 0.5749047300388118,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "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": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "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": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.3333333333333333,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.5121951219512195,
      "mmlu_eval_accuracy_formal_logic": 0.07142857142857142,
      "mmlu_eval_accuracy_global_facts": 0.6,
      "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": 0.7222222222222222,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5116279069767442,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "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": 0.43478260869565216,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "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": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "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": 0.31,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.6285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_law": 0.37058823529411766,
      "mmlu_eval_accuracy_professional_medicine": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_psychology": 0.5507246376811594,
      "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": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.556604993639042,
      "step": 4000
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2511,
      "step": 4001
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3739,
      "step": 4002
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2966,
      "step": 4003
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2985,
      "step": 4004
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2718,
      "step": 4005
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4299,
      "step": 4006
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3513,
      "step": 4007
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2278,
      "step": 4008
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.38,
      "step": 4009
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.346,
      "step": 4010
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.327,
      "step": 4011
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3772,
      "step": 4012
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2162,
      "step": 4013
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4123,
      "step": 4014
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.225,
      "step": 4015
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2453,
      "step": 4016
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3305,
      "step": 4017
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.253,
      "step": 4018
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2269,
      "step": 4019
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3332,
      "step": 4020
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3686,
      "step": 4021
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3449,
      "step": 4022
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.335,
      "step": 4023
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3957,
      "step": 4024
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3882,
      "step": 4025
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2561,
      "step": 4026
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.3203,
      "step": 4027
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.2921,
      "step": 4028
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.1975,
      "step": 4029
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3476,
      "step": 4030
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4438,
      "step": 4031
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3259,
      "step": 4032
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2953,
      "step": 4033
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2612,
      "step": 4034
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2834,
      "step": 4035
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2183,
      "step": 4036
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3501,
      "step": 4037
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3112,
      "step": 4038
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2275,
      "step": 4039
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2608,
      "step": 4040
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4411,
      "step": 4041
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3264,
      "step": 4042
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.262,
      "step": 4043
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3929,
      "step": 4044
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3634,
      "step": 4045
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2881,
      "step": 4046
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2442,
      "step": 4047
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3451,
      "step": 4048
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2434,
      "step": 4049
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3017,
      "step": 4050
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4253,
      "step": 4051
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2858,
      "step": 4052
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3607,
      "step": 4053
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2248,
      "step": 4054
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2942,
      "step": 4055
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3182,
      "step": 4056
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3719,
      "step": 4057
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2416,
      "step": 4058
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.1948,
      "step": 4059
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2581,
      "step": 4060
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3236,
      "step": 4061
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2827,
      "step": 4062
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3476,
      "step": 4063
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3918,
      "step": 4064
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2447,
      "step": 4065
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3334,
      "step": 4066
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3509,
      "step": 4067
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2489,
      "step": 4068
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3752,
      "step": 4069
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3273,
      "step": 4070
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2786,
      "step": 4071
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3207,
      "step": 4072
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2989,
      "step": 4073
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2862,
      "step": 4074
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2797,
      "step": 4075
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.2889,
      "step": 4076
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3141,
      "step": 4077
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3931,
      "step": 4078
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2576,
      "step": 4079
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2727,
      "step": 4080
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2415,
      "step": 4081
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.313,
      "step": 4082
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2498,
      "step": 4083
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3888,
      "step": 4084
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2681,
      "step": 4085
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.319,
      "step": 4086
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2742,
      "step": 4087
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3418,
      "step": 4088
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3304,
      "step": 4089
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3058,
      "step": 4090
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3165,
      "step": 4091
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3269,
      "step": 4092
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3162,
      "step": 4093
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.366,
      "step": 4094
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.4089,
      "step": 4095
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3982,
      "step": 4096
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3924,
      "step": 4097
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2923,
      "step": 4098
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3329,
      "step": 4099
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2476,
      "step": 4100
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2495,
      "step": 4101
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2674,
      "step": 4102
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2556,
      "step": 4103
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.4299,
      "step": 4104
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3088,
      "step": 4105
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3394,
      "step": 4106
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3834,
      "step": 4107
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2251,
      "step": 4108
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2731,
      "step": 4109
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3797,
      "step": 4110
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.407,
      "step": 4111
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.4059,
      "step": 4112
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2296,
      "step": 4113
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2862,
      "step": 4114
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2466,
      "step": 4115
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3892,
      "step": 4116
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3155,
      "step": 4117
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.325,
      "step": 4118
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3727,
      "step": 4119
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3244,
      "step": 4120
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2268,
      "step": 4121
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3045,
      "step": 4122
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2078,
      "step": 4123
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.2684,
      "step": 4124
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.0002,
      "loss": 0.3095,
      "step": 4125
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3388,
      "step": 4126
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2991,
      "step": 4127
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2723,
      "step": 4128
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3084,
      "step": 4129
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2821,
      "step": 4130
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3689,
      "step": 4131
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3523,
      "step": 4132
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.4036,
      "step": 4133
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3271,
      "step": 4134
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2957,
      "step": 4135
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3245,
      "step": 4136
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3721,
      "step": 4137
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2091,
      "step": 4138
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.4912,
      "step": 4139
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3227,
      "step": 4140
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.396,
      "step": 4141
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.4002,
      "step": 4142
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3244,
      "step": 4143
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2469,
      "step": 4144
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.281,
      "step": 4145
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.289,
      "step": 4146
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2664,
      "step": 4147
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.4207,
      "step": 4148
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2952,
      "step": 4149
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2548,
      "step": 4150
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.336,
      "step": 4151
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3848,
      "step": 4152
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3776,
      "step": 4153
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3873,
      "step": 4154
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3608,
      "step": 4155
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3156,
      "step": 4156
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3461,
      "step": 4157
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2279,
      "step": 4158
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.4221,
      "step": 4159
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.4363,
      "step": 4160
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.213,
      "step": 4161
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.4437,
      "step": 4162
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.368,
      "step": 4163
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3834,
      "step": 4164
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2599,
      "step": 4165
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2828,
      "step": 4166
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2364,
      "step": 4167
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2164,
      "step": 4168
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2939,
      "step": 4169
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3429,
      "step": 4170
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.2906,
      "step": 4171
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3155,
      "step": 4172
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0002,
      "loss": 0.3658,
      "step": 4173
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3293,
      "step": 4174
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2212,
      "step": 4175
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2802,
      "step": 4176
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3159,
      "step": 4177
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.299,
      "step": 4178
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3647,
      "step": 4179
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3142,
      "step": 4180
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2654,
      "step": 4181
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.462,
      "step": 4182
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3295,
      "step": 4183
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.371,
      "step": 4184
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2814,
      "step": 4185
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2246,
      "step": 4186
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3828,
      "step": 4187
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2226,
      "step": 4188
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2337,
      "step": 4189
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3329,
      "step": 4190
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.5041,
      "step": 4191
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3456,
      "step": 4192
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.1783,
      "step": 4193
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2627,
      "step": 4194
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3128,
      "step": 4195
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2344,
      "step": 4196
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3885,
      "step": 4197
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3471,
      "step": 4198
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2907,
      "step": 4199
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3081,
      "step": 4200
    },
    {
      "epoch": 0.87,
      "eval_loss": 0.31450411677360535,
      "eval_runtime": 159.4525,
      "eval_samples_per_second": 6.271,
      "eval_steps_per_second": 3.136,
      "step": 4200
    },
    {
      "epoch": 0.87,
      "mmlu_eval_accuracy": 0.5564206571207567,
      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.8125,
      "mmlu_eval_accuracy_business_ethics": 0.7272727272727273,
      "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": 0.18181818181818182,
      "mmlu_eval_accuracy_college_medicine": 0.5,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.5769230769230769,
      "mmlu_eval_accuracy_econometrics": 0.4166666666666667,
      "mmlu_eval_accuracy_electrical_engineering": 0.25,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
      "mmlu_eval_accuracy_formal_logic": 0.07142857142857142,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.5,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "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": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5348837209302325,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
      "mmlu_eval_accuracy_high_school_psychology": 0.8166666666666667,
      "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": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "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": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.6052631578947368,
      "mmlu_eval_accuracy_moral_scenarios": 0.33,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.6571428571428571,
      "mmlu_eval_accuracy_professional_accounting": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_law": 0.3764705882352941,
      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.7727272727272727,
      "mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.5098840920327539,
      "step": 4200
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3938,
      "step": 4201
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.4064,
      "step": 4202
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3106,
      "step": 4203
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2014,
      "step": 4204
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2414,
      "step": 4205
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3054,
      "step": 4206
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2546,
      "step": 4207
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2699,
      "step": 4208
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3362,
      "step": 4209
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2508,
      "step": 4210
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.4899,
      "step": 4211
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2377,
      "step": 4212
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3135,
      "step": 4213
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2716,
      "step": 4214
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3286,
      "step": 4215
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.254,
      "step": 4216
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2663,
      "step": 4217
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2133,
      "step": 4218
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.3082,
      "step": 4219
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2375,
      "step": 4220
    },
    {
      "epoch": 0.87,
      "learning_rate": 0.0002,
      "loss": 0.2779,
      "step": 4221
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.341,
      "step": 4222
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3195,
      "step": 4223
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.4238,
      "step": 4224
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.4401,
      "step": 4225
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2915,
      "step": 4226
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3241,
      "step": 4227
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.1618,
      "step": 4228
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2902,
      "step": 4229
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.4571,
      "step": 4230
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.4122,
      "step": 4231
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3846,
      "step": 4232
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.1833,
      "step": 4233
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2659,
      "step": 4234
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2802,
      "step": 4235
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2113,
      "step": 4236
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3117,
      "step": 4237
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2296,
      "step": 4238
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2834,
      "step": 4239
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2571,
      "step": 4240
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.257,
      "step": 4241
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3197,
      "step": 4242
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2373,
      "step": 4243
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2942,
      "step": 4244
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3757,
      "step": 4245
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3481,
      "step": 4246
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.4528,
      "step": 4247
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.4469,
      "step": 4248
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3921,
      "step": 4249
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3071,
      "step": 4250
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3095,
      "step": 4251
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2604,
      "step": 4252
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3632,
      "step": 4253
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.5205,
      "step": 4254
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.285,
      "step": 4255
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2978,
      "step": 4256
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2525,
      "step": 4257
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2581,
      "step": 4258
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.209,
      "step": 4259
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3455,
      "step": 4260
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3306,
      "step": 4261
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2711,
      "step": 4262
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.236,
      "step": 4263
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.2654,
      "step": 4264
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.1734,
      "step": 4265
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3812,
      "step": 4266
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3109,
      "step": 4267
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.3584,
      "step": 4268
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0002,
      "loss": 0.28,
      "step": 4269
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3086,
      "step": 4270
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3244,
      "step": 4271
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.4048,
      "step": 4272
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2854,
      "step": 4273
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2918,
      "step": 4274
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2073,
      "step": 4275
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2994,
      "step": 4276
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2714,
      "step": 4277
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3321,
      "step": 4278
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2831,
      "step": 4279
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2334,
      "step": 4280
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2608,
      "step": 4281
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2344,
      "step": 4282
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3088,
      "step": 4283
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.391,
      "step": 4284
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2754,
      "step": 4285
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2322,
      "step": 4286
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3012,
      "step": 4287
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2847,
      "step": 4288
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3482,
      "step": 4289
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2307,
      "step": 4290
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3349,
      "step": 4291
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2277,
      "step": 4292
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3276,
      "step": 4293
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3311,
      "step": 4294
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2924,
      "step": 4295
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2469,
      "step": 4296
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.1816,
      "step": 4297
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2245,
      "step": 4298
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3,
      "step": 4299
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3301,
      "step": 4300
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2803,
      "step": 4301
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3835,
      "step": 4302
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.4373,
      "step": 4303
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3173,
      "step": 4304
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.4621,
      "step": 4305
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2681,
      "step": 4306
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2685,
      "step": 4307
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.309,
      "step": 4308
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3256,
      "step": 4309
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2693,
      "step": 4310
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2267,
      "step": 4311
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.196,
      "step": 4312
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.4306,
      "step": 4313
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2904,
      "step": 4314
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2858,
      "step": 4315
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.2932,
      "step": 4316
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3132,
      "step": 4317
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.0002,
      "loss": 0.3351,
      "step": 4318
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.5166,
      "step": 4319
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3021,
      "step": 4320
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3245,
      "step": 4321
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2629,
      "step": 4322
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2865,
      "step": 4323
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2184,
      "step": 4324
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2725,
      "step": 4325
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3192,
      "step": 4326
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2761,
      "step": 4327
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2479,
      "step": 4328
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3642,
      "step": 4329
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2694,
      "step": 4330
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.4094,
      "step": 4331
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2382,
      "step": 4332
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3378,
      "step": 4333
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.4078,
      "step": 4334
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3122,
      "step": 4335
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.233,
      "step": 4336
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2618,
      "step": 4337
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3134,
      "step": 4338
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.1935,
      "step": 4339
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.4391,
      "step": 4340
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2521,
      "step": 4341
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3632,
      "step": 4342
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.291,
      "step": 4343
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.4093,
      "step": 4344
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3112,
      "step": 4345
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3016,
      "step": 4346
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.5323,
      "step": 4347
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2777,
      "step": 4348
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3205,
      "step": 4349
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3274,
      "step": 4350
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2391,
      "step": 4351
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2355,
      "step": 4352
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2302,
      "step": 4353
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3524,
      "step": 4354
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.5637,
      "step": 4355
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.4395,
      "step": 4356
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3226,
      "step": 4357
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.4533,
      "step": 4358
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2961,
      "step": 4359
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3446,
      "step": 4360
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2394,
      "step": 4361
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.1984,
      "step": 4362
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2479,
      "step": 4363
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.3157,
      "step": 4364
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.2984,
      "step": 4365
    },
    {
      "epoch": 0.9,
      "learning_rate": 0.0002,
      "loss": 0.504,
      "step": 4366
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3117,
      "step": 4367
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3756,
      "step": 4368
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3898,
      "step": 4369
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3314,
      "step": 4370
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2849,
      "step": 4371
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2797,
      "step": 4372
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2604,
      "step": 4373
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.4492,
      "step": 4374
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3373,
      "step": 4375
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.4683,
      "step": 4376
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3688,
      "step": 4377
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2816,
      "step": 4378
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.4305,
      "step": 4379
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2053,
      "step": 4380
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2097,
      "step": 4381
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3588,
      "step": 4382
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3617,
      "step": 4383
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3979,
      "step": 4384
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.306,
      "step": 4385
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2627,
      "step": 4386
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2917,
      "step": 4387
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3011,
      "step": 4388
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2217,
      "step": 4389
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3857,
      "step": 4390
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2977,
      "step": 4391
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3169,
      "step": 4392
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3873,
      "step": 4393
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2141,
      "step": 4394
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2809,
      "step": 4395
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2887,
      "step": 4396
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2964,
      "step": 4397
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2756,
      "step": 4398
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.5632,
      "step": 4399
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.493,
      "step": 4400
    },
    {
      "epoch": 0.91,
      "eval_loss": 0.3138442933559418,
      "eval_runtime": 159.2112,
      "eval_samples_per_second": 6.281,
      "eval_steps_per_second": 3.14,
      "step": 4400
    },
    {
      "epoch": 0.91,
      "mmlu_eval_accuracy": 0.5821271781032563,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.8125,
      "mmlu_eval_accuracy_business_ethics": 0.8181818181818182,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
      "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": 0.18181818181818182,
      "mmlu_eval_accuracy_college_medicine": 0.5,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5769230769230769,
      "mmlu_eval_accuracy_econometrics": 0.3333333333333333,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.6,
      "mmlu_eval_accuracy_high_school_biology": 0.5,
      "mmlu_eval_accuracy_high_school_chemistry": 0.5454545454545454,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.35294117647058826,
      "mmlu_eval_accuracy_high_school_psychology": 0.7833333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.8260869565217391,
      "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": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7558139534883721,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.28,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_law": 0.3764705882352941,
      "mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_psychology": 0.5072463768115942,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6666666666666666,
      "mmlu_eval_accuracy_sociology": 0.7727272727272727,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.6535655940948517,
      "step": 4400
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3397,
      "step": 4401
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.4102,
      "step": 4402
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.4176,
      "step": 4403
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3388,
      "step": 4404
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2793,
      "step": 4405
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2932,
      "step": 4406
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3467,
      "step": 4407
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3086,
      "step": 4408
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.347,
      "step": 4409
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3067,
      "step": 4410
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.3822,
      "step": 4411
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.27,
      "step": 4412
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2919,
      "step": 4413
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.0002,
      "loss": 0.2575,
      "step": 4414
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3353,
      "step": 4415
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.4506,
      "step": 4416
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3893,
      "step": 4417
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2847,
      "step": 4418
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2285,
      "step": 4419
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2493,
      "step": 4420
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3222,
      "step": 4421
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3773,
      "step": 4422
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3301,
      "step": 4423
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2668,
      "step": 4424
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2833,
      "step": 4425
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2871,
      "step": 4426
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2186,
      "step": 4427
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.229,
      "step": 4428
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2591,
      "step": 4429
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2245,
      "step": 4430
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.4238,
      "step": 4431
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.322,
      "step": 4432
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2345,
      "step": 4433
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2886,
      "step": 4434
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2136,
      "step": 4435
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2269,
      "step": 4436
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3385,
      "step": 4437
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2938,
      "step": 4438
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3775,
      "step": 4439
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3544,
      "step": 4440
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.4667,
      "step": 4441
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2649,
      "step": 4442
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.5981,
      "step": 4443
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2908,
      "step": 4444
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2977,
      "step": 4445
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2795,
      "step": 4446
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2751,
      "step": 4447
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3676,
      "step": 4448
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2915,
      "step": 4449
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2938,
      "step": 4450
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3191,
      "step": 4451
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3657,
      "step": 4452
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.4714,
      "step": 4453
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3477,
      "step": 4454
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.4115,
      "step": 4455
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3107,
      "step": 4456
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3121,
      "step": 4457
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.274,
      "step": 4458
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3401,
      "step": 4459
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.2835,
      "step": 4460
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3079,
      "step": 4461
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0002,
      "loss": 0.3719,
      "step": 4462
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2898,
      "step": 4463
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.4736,
      "step": 4464
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3322,
      "step": 4465
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2589,
      "step": 4466
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2526,
      "step": 4467
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3673,
      "step": 4468
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3774,
      "step": 4469
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3796,
      "step": 4470
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3536,
      "step": 4471
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2764,
      "step": 4472
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3602,
      "step": 4473
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2452,
      "step": 4474
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2528,
      "step": 4475
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3751,
      "step": 4476
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3042,
      "step": 4477
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2539,
      "step": 4478
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3035,
      "step": 4479
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3902,
      "step": 4480
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3444,
      "step": 4481
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2559,
      "step": 4482
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2364,
      "step": 4483
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.4398,
      "step": 4484
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3272,
      "step": 4485
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.326,
      "step": 4486
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2889,
      "step": 4487
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3295,
      "step": 4488
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3204,
      "step": 4489
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2198,
      "step": 4490
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.304,
      "step": 4491
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2491,
      "step": 4492
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.5241,
      "step": 4493
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.227,
      "step": 4494
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3723,
      "step": 4495
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2917,
      "step": 4496
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3271,
      "step": 4497
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2547,
      "step": 4498
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2113,
      "step": 4499
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2852,
      "step": 4500
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.5149,
      "step": 4501
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.5134,
      "step": 4502
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3234,
      "step": 4503
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2367,
      "step": 4504
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2201,
      "step": 4505
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.254,
      "step": 4506
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.4319,
      "step": 4507
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3798,
      "step": 4508
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2222,
      "step": 4509
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.2236,
      "step": 4510
    },
    {
      "epoch": 0.93,
      "learning_rate": 0.0002,
      "loss": 0.3105,
      "step": 4511
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.228,
      "step": 4512
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.1945,
      "step": 4513
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3753,
      "step": 4514
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.4456,
      "step": 4515
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3107,
      "step": 4516
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3138,
      "step": 4517
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3182,
      "step": 4518
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3381,
      "step": 4519
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2091,
      "step": 4520
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3556,
      "step": 4521
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3699,
      "step": 4522
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3832,
      "step": 4523
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3173,
      "step": 4524
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3728,
      "step": 4525
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2622,
      "step": 4526
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2717,
      "step": 4527
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3204,
      "step": 4528
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3687,
      "step": 4529
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3021,
      "step": 4530
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.346,
      "step": 4531
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2556,
      "step": 4532
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2803,
      "step": 4533
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2904,
      "step": 4534
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2968,
      "step": 4535
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.566,
      "step": 4536
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2857,
      "step": 4537
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.4381,
      "step": 4538
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3964,
      "step": 4539
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3105,
      "step": 4540
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2868,
      "step": 4541
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.4253,
      "step": 4542
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3076,
      "step": 4543
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.234,
      "step": 4544
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.7355,
      "step": 4545
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3452,
      "step": 4546
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3197,
      "step": 4547
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3547,
      "step": 4548
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.4245,
      "step": 4549
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2516,
      "step": 4550
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3785,
      "step": 4551
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2709,
      "step": 4552
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3248,
      "step": 4553
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.3389,
      "step": 4554
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2407,
      "step": 4555
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2751,
      "step": 4556
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2388,
      "step": 4557
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.2626,
      "step": 4558
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.0002,
      "loss": 0.389,
      "step": 4559
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2979,
      "step": 4560
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3124,
      "step": 4561
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3031,
      "step": 4562
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2945,
      "step": 4563
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.305,
      "step": 4564
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3726,
      "step": 4565
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2512,
      "step": 4566
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2881,
      "step": 4567
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3255,
      "step": 4568
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3813,
      "step": 4569
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3831,
      "step": 4570
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.5496,
      "step": 4571
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2622,
      "step": 4572
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3333,
      "step": 4573
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2794,
      "step": 4574
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3198,
      "step": 4575
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2841,
      "step": 4576
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.1758,
      "step": 4577
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3296,
      "step": 4578
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2906,
      "step": 4579
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3864,
      "step": 4580
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.4114,
      "step": 4581
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3336,
      "step": 4582
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3544,
      "step": 4583
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.299,
      "step": 4584
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3022,
      "step": 4585
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2987,
      "step": 4586
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3324,
      "step": 4587
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.211,
      "step": 4588
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.4226,
      "step": 4589
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2545,
      "step": 4590
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2186,
      "step": 4591
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.4168,
      "step": 4592
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3246,
      "step": 4593
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3125,
      "step": 4594
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.243,
      "step": 4595
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2766,
      "step": 4596
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3763,
      "step": 4597
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2487,
      "step": 4598
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.3593,
      "step": 4599
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2216,
      "step": 4600
    },
    {
      "epoch": 0.95,
      "eval_loss": 0.3118540048599243,
      "eval_runtime": 159.3089,
      "eval_samples_per_second": 6.277,
      "eval_steps_per_second": 3.139,
      "step": 4600
    },
    {
      "epoch": 0.95,
      "mmlu_eval_accuracy": 0.5574952168735646,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.7272727272727273,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
      "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": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.5,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.25,
      "mmlu_eval_accuracy_electrical_engineering": 0.3125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.53125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "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": 0.6190476190476191,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5581395348837209,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "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": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.8695652173913043,
      "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": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "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": 0.31,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.6764705882352942,
      "mmlu_eval_accuracy_prehistory": 0.6285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_law": 0.38235294117647056,
      "mmlu_eval_accuracy_professional_medicine": 0.5806451612903226,
      "mmlu_eval_accuracy_professional_psychology": 0.5217391304347826,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.6666666666666666,
      "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": 1.7436769160371537,
      "step": 4600
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.4778,
      "step": 4601
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.1787,
      "step": 4602
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2966,
      "step": 4603
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.4486,
      "step": 4604
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2518,
      "step": 4605
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.4142,
      "step": 4606
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.0002,
      "loss": 0.2826,
      "step": 4607
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2391,
      "step": 4608
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.4062,
      "step": 4609
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.4047,
      "step": 4610
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2228,
      "step": 4611
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.4341,
      "step": 4612
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2884,
      "step": 4613
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2542,
      "step": 4614
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3603,
      "step": 4615
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3755,
      "step": 4616
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2676,
      "step": 4617
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3809,
      "step": 4618
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3425,
      "step": 4619
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2908,
      "step": 4620
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.1718,
      "step": 4621
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3226,
      "step": 4622
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.4039,
      "step": 4623
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3205,
      "step": 4624
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3223,
      "step": 4625
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.4104,
      "step": 4626
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3185,
      "step": 4627
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.266,
      "step": 4628
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2748,
      "step": 4629
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.386,
      "step": 4630
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2003,
      "step": 4631
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3496,
      "step": 4632
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3185,
      "step": 4633
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2522,
      "step": 4634
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3488,
      "step": 4635
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2576,
      "step": 4636
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3477,
      "step": 4637
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2393,
      "step": 4638
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2408,
      "step": 4639
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3273,
      "step": 4640
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2207,
      "step": 4641
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3776,
      "step": 4642
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2043,
      "step": 4643
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2335,
      "step": 4644
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.331,
      "step": 4645
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3033,
      "step": 4646
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3253,
      "step": 4647
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2761,
      "step": 4648
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3427,
      "step": 4649
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3316,
      "step": 4650
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2696,
      "step": 4651
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3265,
      "step": 4652
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.3284,
      "step": 4653
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.244,
      "step": 4654
    },
    {
      "epoch": 0.96,
      "learning_rate": 0.0002,
      "loss": 0.2296,
      "step": 4655
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3142,
      "step": 4656
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3906,
      "step": 4657
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3516,
      "step": 4658
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2998,
      "step": 4659
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2512,
      "step": 4660
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.1772,
      "step": 4661
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2605,
      "step": 4662
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.4928,
      "step": 4663
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2736,
      "step": 4664
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.5092,
      "step": 4665
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3062,
      "step": 4666
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3258,
      "step": 4667
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2914,
      "step": 4668
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3441,
      "step": 4669
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2239,
      "step": 4670
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2479,
      "step": 4671
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2702,
      "step": 4672
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2245,
      "step": 4673
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2217,
      "step": 4674
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.316,
      "step": 4675
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3054,
      "step": 4676
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2859,
      "step": 4677
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2932,
      "step": 4678
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2495,
      "step": 4679
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3745,
      "step": 4680
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3248,
      "step": 4681
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2887,
      "step": 4682
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3769,
      "step": 4683
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.1881,
      "step": 4684
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3459,
      "step": 4685
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3311,
      "step": 4686
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3324,
      "step": 4687
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.3764,
      "step": 4688
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2433,
      "step": 4689
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.1588,
      "step": 4690
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.4959,
      "step": 4691
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2339,
      "step": 4692
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.263,
      "step": 4693
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2704,
      "step": 4694
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.347,
      "step": 4695
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2353,
      "step": 4696
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2775,
      "step": 4697
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.224,
      "step": 4698
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2738,
      "step": 4699
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.44,
      "step": 4700
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.4081,
      "step": 4701
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.4042,
      "step": 4702
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2525,
      "step": 4703
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0002,
      "loss": 0.2222,
      "step": 4704
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3454,
      "step": 4705
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2616,
      "step": 4706
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2329,
      "step": 4707
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.233,
      "step": 4708
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3532,
      "step": 4709
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3316,
      "step": 4710
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2742,
      "step": 4711
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2804,
      "step": 4712
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3173,
      "step": 4713
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3143,
      "step": 4714
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2724,
      "step": 4715
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2791,
      "step": 4716
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3739,
      "step": 4717
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3251,
      "step": 4718
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3203,
      "step": 4719
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3463,
      "step": 4720
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3146,
      "step": 4721
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.4241,
      "step": 4722
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.5414,
      "step": 4723
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3389,
      "step": 4724
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2643,
      "step": 4725
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2321,
      "step": 4726
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2245,
      "step": 4727
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.304,
      "step": 4728
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3147,
      "step": 4729
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.285,
      "step": 4730
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3726,
      "step": 4731
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.485,
      "step": 4732
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2439,
      "step": 4733
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3095,
      "step": 4734
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3531,
      "step": 4735
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3428,
      "step": 4736
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3789,
      "step": 4737
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.1901,
      "step": 4738
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2763,
      "step": 4739
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2907,
      "step": 4740
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2882,
      "step": 4741
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3172,
      "step": 4742
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.4274,
      "step": 4743
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.4204,
      "step": 4744
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2334,
      "step": 4745
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2531,
      "step": 4746
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3106,
      "step": 4747
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2416,
      "step": 4748
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2781,
      "step": 4749
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3015,
      "step": 4750
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.3569,
      "step": 4751
    },
    {
      "epoch": 0.98,
      "learning_rate": 0.0002,
      "loss": 0.2838,
      "step": 4752
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2764,
      "step": 4753
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3091,
      "step": 4754
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.459,
      "step": 4755
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3104,
      "step": 4756
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2128,
      "step": 4757
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2762,
      "step": 4758
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3247,
      "step": 4759
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2202,
      "step": 4760
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3228,
      "step": 4761
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.1863,
      "step": 4762
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.4038,
      "step": 4763
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3318,
      "step": 4764
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3202,
      "step": 4765
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2654,
      "step": 4766
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3936,
      "step": 4767
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2333,
      "step": 4768
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.4104,
      "step": 4769
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.1972,
      "step": 4770
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.4406,
      "step": 4771
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2725,
      "step": 4772
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2712,
      "step": 4773
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2892,
      "step": 4774
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2951,
      "step": 4775
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.415,
      "step": 4776
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2684,
      "step": 4777
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2431,
      "step": 4778
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3071,
      "step": 4779
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3311,
      "step": 4780
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3146,
      "step": 4781
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2383,
      "step": 4782
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.4813,
      "step": 4783
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3578,
      "step": 4784
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.5106,
      "step": 4785
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3456,
      "step": 4786
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2684,
      "step": 4787
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2078,
      "step": 4788
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2927,
      "step": 4789
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2868,
      "step": 4790
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3041,
      "step": 4791
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3079,
      "step": 4792
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3111,
      "step": 4793
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.5282,
      "step": 4794
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2699,
      "step": 4795
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2794,
      "step": 4796
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3439,
      "step": 4797
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.246,
      "step": 4798
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.3042,
      "step": 4799
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0002,
      "loss": 0.2984,
      "step": 4800
    },
    {
      "epoch": 0.99,
      "eval_loss": 0.31124767661094666,
      "eval_runtime": 159.3971,
      "eval_samples_per_second": 6.274,
      "eval_steps_per_second": 3.137,
      "step": 4800
    },
    {
      "epoch": 0.99,
      "mmlu_eval_accuracy": 0.5518233366256161,
      "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": 0.5517241379310345,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
      "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": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.4166666666666667,
      "mmlu_eval_accuracy_electrical_engineering": 0.3125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.07142857142857142,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "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": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5581395348837209,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6923076923076923,
      "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
      "mmlu_eval_accuracy_high_school_psychology": 0.7666666666666667,
      "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": 0.8260869565217391,
      "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": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.8181818181818182,
      "mmlu_eval_accuracy_miscellaneous": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.3,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.6764705882352942,
      "mmlu_eval_accuracy_prehistory": 0.6857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_law": 0.3764705882352941,
      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.6296296296296297,
      "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": 1.791161036133455,
      "step": 4800
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3665,
      "step": 4801
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.4054,
      "step": 4802
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3715,
      "step": 4803
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3313,
      "step": 4804
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.1939,
      "step": 4805
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3726,
      "step": 4806
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3088,
      "step": 4807
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2758,
      "step": 4808
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2921,
      "step": 4809
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.4074,
      "step": 4810
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2846,
      "step": 4811
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2281,
      "step": 4812
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2077,
      "step": 4813
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2513,
      "step": 4814
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2335,
      "step": 4815
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3533,
      "step": 4816
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3557,
      "step": 4817
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3106,
      "step": 4818
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.4176,
      "step": 4819
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3522,
      "step": 4820
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3711,
      "step": 4821
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3497,
      "step": 4822
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2397,
      "step": 4823
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.5627,
      "step": 4824
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.1757,
      "step": 4825
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2391,
      "step": 4826
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.1732,
      "step": 4827
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2168,
      "step": 4828
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2339,
      "step": 4829
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.1769,
      "step": 4830
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.4001,
      "step": 4831
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2032,
      "step": 4832
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.241,
      "step": 4833
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2834,
      "step": 4834
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2443,
      "step": 4835
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3429,
      "step": 4836
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.202,
      "step": 4837
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2445,
      "step": 4838
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.24,
      "step": 4839
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2276,
      "step": 4840
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2085,
      "step": 4841
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.1807,
      "step": 4842
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3168,
      "step": 4843
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.3777,
      "step": 4844
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2212,
      "step": 4845
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.1805,
      "step": 4846
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.2252,
      "step": 4847
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.0002,
      "loss": 0.1637,
      "step": 4848
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.3074,
      "step": 4849
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2155,
      "step": 4850
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.1775,
      "step": 4851
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2571,
      "step": 4852
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.3516,
      "step": 4853
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.1896,
      "step": 4854
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.1561,
      "step": 4855
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.306,
      "step": 4856
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2018,
      "step": 4857
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.3257,
      "step": 4858
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.427,
      "step": 4859
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.1957,
      "step": 4860
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2029,
      "step": 4861
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2075,
      "step": 4862
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2226,
      "step": 4863
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.3459,
      "step": 4864
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2321,
      "step": 4865
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.1901,
      "step": 4866
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.3737,
      "step": 4867
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2139,
      "step": 4868
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2219,
      "step": 4869
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2387,
      "step": 4870
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.3298,
      "step": 4871
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.3362,
      "step": 4872
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.1693,
      "step": 4873
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2947,
      "step": 4874
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2253,
      "step": 4875
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.1593,
      "step": 4876
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2169,
      "step": 4877
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2123,
      "step": 4878
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2318,
      "step": 4879
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2162,
      "step": 4880
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2047,
      "step": 4881
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.1929,
      "step": 4882
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2499,
      "step": 4883
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2244,
      "step": 4884
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2532,
      "step": 4885
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2264,
      "step": 4886
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2802,
      "step": 4887
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2634,
      "step": 4888
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2409,
      "step": 4889
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.4498,
      "step": 4890
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.3064,
      "step": 4891
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2196,
      "step": 4892
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2147,
      "step": 4893
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2068,
      "step": 4894
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.3232,
      "step": 4895
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2558,
      "step": 4896
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0002,
      "loss": 0.2375,
      "step": 4897
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.1793,
      "step": 4898
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.1742,
      "step": 4899
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.3233,
      "step": 4900
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2361,
      "step": 4901
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2264,
      "step": 4902
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2768,
      "step": 4903
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2108,
      "step": 4904
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2887,
      "step": 4905
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.3019,
      "step": 4906
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2658,
      "step": 4907
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2898,
      "step": 4908
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.3816,
      "step": 4909
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2248,
      "step": 4910
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2811,
      "step": 4911
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.1991,
      "step": 4912
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.1885,
      "step": 4913
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2447,
      "step": 4914
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.4129,
      "step": 4915
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.1919,
      "step": 4916
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2014,
      "step": 4917
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.1926,
      "step": 4918
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.1915,
      "step": 4919
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2708,
      "step": 4920
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2358,
      "step": 4921
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2067,
      "step": 4922
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2411,
      "step": 4923
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2857,
      "step": 4924
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2245,
      "step": 4925
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2688,
      "step": 4926
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2057,
      "step": 4927
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.3266,
      "step": 4928
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.1919,
      "step": 4929
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2012,
      "step": 4930
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2182,
      "step": 4931
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.3538,
      "step": 4932
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2318,
      "step": 4933
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2495,
      "step": 4934
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.1973,
      "step": 4935
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2104,
      "step": 4936
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.1978,
      "step": 4937
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.3077,
      "step": 4938
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.1889,
      "step": 4939
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2143,
      "step": 4940
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2606,
      "step": 4941
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2569,
      "step": 4942
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2174,
      "step": 4943
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2586,
      "step": 4944
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.0002,
      "loss": 0.2035,
      "step": 4945
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2103,
      "step": 4946
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2657,
      "step": 4947
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.3771,
      "step": 4948
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2573,
      "step": 4949
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2142,
      "step": 4950
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1993,
      "step": 4951
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1859,
      "step": 4952
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2687,
      "step": 4953
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1799,
      "step": 4954
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2724,
      "step": 4955
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.246,
      "step": 4956
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.228,
      "step": 4957
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2614,
      "step": 4958
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.3018,
      "step": 4959
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1688,
      "step": 4960
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2648,
      "step": 4961
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1887,
      "step": 4962
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2274,
      "step": 4963
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2869,
      "step": 4964
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.213,
      "step": 4965
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1798,
      "step": 4966
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2027,
      "step": 4967
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2366,
      "step": 4968
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1862,
      "step": 4969
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2296,
      "step": 4970
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.3789,
      "step": 4971
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2818,
      "step": 4972
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2291,
      "step": 4973
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.233,
      "step": 4974
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.248,
      "step": 4975
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2541,
      "step": 4976
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1687,
      "step": 4977
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1886,
      "step": 4978
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2112,
      "step": 4979
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2984,
      "step": 4980
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2231,
      "step": 4981
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.365,
      "step": 4982
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1729,
      "step": 4983
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1866,
      "step": 4984
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2372,
      "step": 4985
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1504,
      "step": 4986
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.198,
      "step": 4987
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1927,
      "step": 4988
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.183,
      "step": 4989
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2215,
      "step": 4990
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.1244,
      "step": 4991
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.32,
      "step": 4992
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0002,
      "loss": 0.2331,
      "step": 4993
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2213,
      "step": 4994
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.3001,
      "step": 4995
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1961,
      "step": 4996
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2275,
      "step": 4997
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1879,
      "step": 4998
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2061,
      "step": 4999
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1764,
      "step": 5000
    },
    {
      "epoch": 1.04,
      "eval_loss": 0.3192247152328491,
      "eval_runtime": 159.3242,
      "eval_samples_per_second": 6.277,
      "eval_steps_per_second": 3.138,
      "step": 5000
    },
    {
      "epoch": 1.04,
      "mmlu_eval_accuracy": 0.5600113979051913,
      "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": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.5,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "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": 0.07142857142857142,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.5,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5348837209302325,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6923076923076923,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8,
      "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": 0.782608695652174,
      "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": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7558139534883721,
      "mmlu_eval_accuracy_moral_disputes": 0.6052631578947368,
      "mmlu_eval_accuracy_moral_scenarios": 0.29,
      "mmlu_eval_accuracy_nutrition": 0.6060606060606061,
      "mmlu_eval_accuracy_philosophy": 0.6470588235294118,
      "mmlu_eval_accuracy_prehistory": 0.6571428571428571,
      "mmlu_eval_accuracy_professional_accounting": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_law": 0.3764705882352941,
      "mmlu_eval_accuracy_professional_medicine": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.7727272727272727,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.7368671817038763,
      "step": 5000
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1452,
      "step": 5001
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1581,
      "step": 5002
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1639,
      "step": 5003
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.213,
      "step": 5004
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.3496,
      "step": 5005
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2344,
      "step": 5006
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2242,
      "step": 5007
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2284,
      "step": 5008
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1776,
      "step": 5009
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2016,
      "step": 5010
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2516,
      "step": 5011
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.14,
      "step": 5012
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2933,
      "step": 5013
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1692,
      "step": 5014
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.236,
      "step": 5015
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2031,
      "step": 5016
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.3539,
      "step": 5017
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.3255,
      "step": 5018
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.167,
      "step": 5019
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2188,
      "step": 5020
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2512,
      "step": 5021
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.3104,
      "step": 5022
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.237,
      "step": 5023
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2494,
      "step": 5024
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2867,
      "step": 5025
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2316,
      "step": 5026
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1789,
      "step": 5027
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2491,
      "step": 5028
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2952,
      "step": 5029
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2557,
      "step": 5030
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1663,
      "step": 5031
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1744,
      "step": 5032
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.3549,
      "step": 5033
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2682,
      "step": 5034
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1658,
      "step": 5035
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2132,
      "step": 5036
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2446,
      "step": 5037
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.1783,
      "step": 5038
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2268,
      "step": 5039
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.2129,
      "step": 5040
    },
    {
      "epoch": 1.04,
      "learning_rate": 0.0002,
      "loss": 0.4271,
      "step": 5041
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2371,
      "step": 5042
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.288,
      "step": 5043
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2285,
      "step": 5044
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.1929,
      "step": 5045
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2163,
      "step": 5046
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2053,
      "step": 5047
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2123,
      "step": 5048
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.3038,
      "step": 5049
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.3386,
      "step": 5050
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.1748,
      "step": 5051
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2606,
      "step": 5052
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2184,
      "step": 5053
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2285,
      "step": 5054
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2619,
      "step": 5055
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2509,
      "step": 5056
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.4831,
      "step": 5057
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2472,
      "step": 5058
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.4239,
      "step": 5059
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2232,
      "step": 5060
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2791,
      "step": 5061
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.1822,
      "step": 5062
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2275,
      "step": 5063
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2536,
      "step": 5064
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2228,
      "step": 5065
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.1575,
      "step": 5066
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2705,
      "step": 5067
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2351,
      "step": 5068
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.1837,
      "step": 5069
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.205,
      "step": 5070
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2185,
      "step": 5071
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.3589,
      "step": 5072
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2366,
      "step": 5073
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2293,
      "step": 5074
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.1959,
      "step": 5075
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.1875,
      "step": 5076
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.4742,
      "step": 5077
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2419,
      "step": 5078
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.1958,
      "step": 5079
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.182,
      "step": 5080
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2408,
      "step": 5081
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2132,
      "step": 5082
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.1911,
      "step": 5083
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.226,
      "step": 5084
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2265,
      "step": 5085
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2464,
      "step": 5086
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2246,
      "step": 5087
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2813,
      "step": 5088
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.2178,
      "step": 5089
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.0002,
      "loss": 0.1929,
      "step": 5090
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1325,
      "step": 5091
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2023,
      "step": 5092
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1807,
      "step": 5093
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2236,
      "step": 5094
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2175,
      "step": 5095
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2328,
      "step": 5096
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.4132,
      "step": 5097
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2696,
      "step": 5098
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1803,
      "step": 5099
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.3756,
      "step": 5100
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2142,
      "step": 5101
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2356,
      "step": 5102
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2311,
      "step": 5103
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.286,
      "step": 5104
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2191,
      "step": 5105
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2143,
      "step": 5106
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1936,
      "step": 5107
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2791,
      "step": 5108
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1881,
      "step": 5109
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2608,
      "step": 5110
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.307,
      "step": 5111
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1901,
      "step": 5112
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1807,
      "step": 5113
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1611,
      "step": 5114
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2428,
      "step": 5115
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2062,
      "step": 5116
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2457,
      "step": 5117
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2768,
      "step": 5118
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1864,
      "step": 5119
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.3043,
      "step": 5120
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2894,
      "step": 5121
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2717,
      "step": 5122
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2559,
      "step": 5123
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1643,
      "step": 5124
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2188,
      "step": 5125
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1984,
      "step": 5126
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1805,
      "step": 5127
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2253,
      "step": 5128
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1374,
      "step": 5129
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2024,
      "step": 5130
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2465,
      "step": 5131
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2276,
      "step": 5132
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2289,
      "step": 5133
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2189,
      "step": 5134
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2781,
      "step": 5135
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2064,
      "step": 5136
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.1647,
      "step": 5137
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.0002,
      "loss": 0.2509,
      "step": 5138
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2912,
      "step": 5139
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2208,
      "step": 5140
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2669,
      "step": 5141
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.333,
      "step": 5142
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.3769,
      "step": 5143
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.3608,
      "step": 5144
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2324,
      "step": 5145
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2537,
      "step": 5146
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2161,
      "step": 5147
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2152,
      "step": 5148
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.1905,
      "step": 5149
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2713,
      "step": 5150
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2923,
      "step": 5151
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2341,
      "step": 5152
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.1871,
      "step": 5153
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2342,
      "step": 5154
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2661,
      "step": 5155
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2051,
      "step": 5156
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.1685,
      "step": 5157
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.1599,
      "step": 5158
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.1745,
      "step": 5159
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.214,
      "step": 5160
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2319,
      "step": 5161
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.188,
      "step": 5162
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.1918,
      "step": 5163
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2639,
      "step": 5164
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.1954,
      "step": 5165
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2233,
      "step": 5166
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.333,
      "step": 5167
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.3205,
      "step": 5168
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.219,
      "step": 5169
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2467,
      "step": 5170
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.19,
      "step": 5171
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.1522,
      "step": 5172
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.3616,
      "step": 5173
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2985,
      "step": 5174
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2176,
      "step": 5175
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.1546,
      "step": 5176
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2006,
      "step": 5177
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.3143,
      "step": 5178
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.4332,
      "step": 5179
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.3678,
      "step": 5180
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2061,
      "step": 5181
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2421,
      "step": 5182
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.2279,
      "step": 5183
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.1808,
      "step": 5184
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.1979,
      "step": 5185
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.0002,
      "loss": 0.1722,
      "step": 5186
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.278,
      "step": 5187
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.1908,
      "step": 5188
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.1936,
      "step": 5189
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2519,
      "step": 5190
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2321,
      "step": 5191
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2219,
      "step": 5192
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.1894,
      "step": 5193
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.3611,
      "step": 5194
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2674,
      "step": 5195
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.197,
      "step": 5196
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.1528,
      "step": 5197
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2607,
      "step": 5198
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.3547,
      "step": 5199
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.3177,
      "step": 5200
    },
    {
      "epoch": 1.08,
      "eval_loss": 0.3207312524318695,
      "eval_runtime": 159.4097,
      "eval_samples_per_second": 6.273,
      "eval_steps_per_second": 3.137,
      "step": 5200
    },
    {
      "epoch": 1.08,
      "mmlu_eval_accuracy": 0.5629467202030355,
      "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": 0.5,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.2727272727272727,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4878048780487805,
      "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": 0.5555555555555556,
      "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": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5581395348837209,
      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.7307692307692307,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.8,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "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": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "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": 0.33,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.6176470588235294,
      "mmlu_eval_accuracy_prehistory": 0.5428571428571428,
      "mmlu_eval_accuracy_professional_accounting": 0.3870967741935484,
      "mmlu_eval_accuracy_professional_law": 0.38235294117647056,
      "mmlu_eval_accuracy_professional_medicine": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_psychology": 0.5507246376811594,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.8181818181818182,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 1.7740337621449802,
      "step": 5200
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2484,
      "step": 5201
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.254,
      "step": 5202
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.3065,
      "step": 5203
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.3022,
      "step": 5204
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2419,
      "step": 5205
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.1842,
      "step": 5206
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.1272,
      "step": 5207
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2443,
      "step": 5208
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2231,
      "step": 5209
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.1985,
      "step": 5210
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2156,
      "step": 5211
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2808,
      "step": 5212
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2793,
      "step": 5213
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2162,
      "step": 5214
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2189,
      "step": 5215
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.1772,
      "step": 5216
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2181,
      "step": 5217
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2846,
      "step": 5218
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.1828,
      "step": 5219
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.3429,
      "step": 5220
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.1703,
      "step": 5221
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2424,
      "step": 5222
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.1738,
      "step": 5223
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2781,
      "step": 5224
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2746,
      "step": 5225
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.186,
      "step": 5226
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.254,
      "step": 5227
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.3285,
      "step": 5228
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.3265,
      "step": 5229
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2447,
      "step": 5230
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.1756,
      "step": 5231
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.2683,
      "step": 5232
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.3047,
      "step": 5233
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0002,
      "loss": 0.166,
      "step": 5234
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2506,
      "step": 5235
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.3159,
      "step": 5236
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1786,
      "step": 5237
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1771,
      "step": 5238
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2921,
      "step": 5239
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1952,
      "step": 5240
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1599,
      "step": 5241
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2401,
      "step": 5242
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.191,
      "step": 5243
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2554,
      "step": 5244
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.4162,
      "step": 5245
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2541,
      "step": 5246
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2977,
      "step": 5247
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2132,
      "step": 5248
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.25,
      "step": 5249
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1817,
      "step": 5250
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.3326,
      "step": 5251
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1377,
      "step": 5252
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2582,
      "step": 5253
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2343,
      "step": 5254
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.311,
      "step": 5255
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1979,
      "step": 5256
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.3007,
      "step": 5257
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.212,
      "step": 5258
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.3082,
      "step": 5259
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2266,
      "step": 5260
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2493,
      "step": 5261
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1862,
      "step": 5262
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.221,
      "step": 5263
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1981,
      "step": 5264
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1821,
      "step": 5265
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2677,
      "step": 5266
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2033,
      "step": 5267
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2805,
      "step": 5268
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2129,
      "step": 5269
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.3145,
      "step": 5270
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2046,
      "step": 5271
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2886,
      "step": 5272
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2593,
      "step": 5273
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1864,
      "step": 5274
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1848,
      "step": 5275
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2869,
      "step": 5276
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.246,
      "step": 5277
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2097,
      "step": 5278
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2806,
      "step": 5279
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1958,
      "step": 5280
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.2848,
      "step": 5281
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.1419,
      "step": 5282
    },
    {
      "epoch": 1.09,
      "learning_rate": 0.0002,
      "loss": 0.196,
      "step": 5283
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2415,
      "step": 5284
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.263,
      "step": 5285
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.1858,
      "step": 5286
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2342,
      "step": 5287
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2168,
      "step": 5288
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2034,
      "step": 5289
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.1811,
      "step": 5290
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.242,
      "step": 5291
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2646,
      "step": 5292
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.1622,
      "step": 5293
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2339,
      "step": 5294
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2036,
      "step": 5295
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2271,
      "step": 5296
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.3277,
      "step": 5297
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2496,
      "step": 5298
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2031,
      "step": 5299
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2119,
      "step": 5300
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.1938,
      "step": 5301
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2311,
      "step": 5302
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.204,
      "step": 5303
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2329,
      "step": 5304
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.1746,
      "step": 5305
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2245,
      "step": 5306
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2018,
      "step": 5307
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.1861,
      "step": 5308
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.3406,
      "step": 5309
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2979,
      "step": 5310
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2529,
      "step": 5311
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2682,
      "step": 5312
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.1758,
      "step": 5313
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.3156,
      "step": 5314
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2878,
      "step": 5315
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2591,
      "step": 5316
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2682,
      "step": 5317
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2054,
      "step": 5318
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.1771,
      "step": 5319
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.1867,
      "step": 5320
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2419,
      "step": 5321
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2631,
      "step": 5322
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2171,
      "step": 5323
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.1918,
      "step": 5324
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2165,
      "step": 5325
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2647,
      "step": 5326
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.1666,
      "step": 5327
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2055,
      "step": 5328
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2336,
      "step": 5329
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.3351,
      "step": 5330
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.0002,
      "loss": 0.2674,
      "step": 5331
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2446,
      "step": 5332
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2501,
      "step": 5333
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2039,
      "step": 5334
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.1856,
      "step": 5335
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.3562,
      "step": 5336
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2278,
      "step": 5337
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2589,
      "step": 5338
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.1336,
      "step": 5339
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2553,
      "step": 5340
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2586,
      "step": 5341
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2411,
      "step": 5342
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2139,
      "step": 5343
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.3085,
      "step": 5344
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2558,
      "step": 5345
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.1444,
      "step": 5346
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.3393,
      "step": 5347
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.3199,
      "step": 5348
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2736,
      "step": 5349
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2116,
      "step": 5350
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2231,
      "step": 5351
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2069,
      "step": 5352
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2072,
      "step": 5353
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2017,
      "step": 5354
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.301,
      "step": 5355
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2603,
      "step": 5356
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.3404,
      "step": 5357
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.4947,
      "step": 5358
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2731,
      "step": 5359
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.192,
      "step": 5360
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2293,
      "step": 5361
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.1692,
      "step": 5362
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2437,
      "step": 5363
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2676,
      "step": 5364
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.1788,
      "step": 5365
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2151,
      "step": 5366
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.3716,
      "step": 5367
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2409,
      "step": 5368
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2843,
      "step": 5369
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2347,
      "step": 5370
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.1836,
      "step": 5371
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2002,
      "step": 5372
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2187,
      "step": 5373
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.3076,
      "step": 5374
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2547,
      "step": 5375
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2534,
      "step": 5376
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.2695,
      "step": 5377
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.3366,
      "step": 5378
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.0002,
      "loss": 0.3366,
      "step": 5379
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2852,
      "step": 5380
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2784,
      "step": 5381
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.3393,
      "step": 5382
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1648,
      "step": 5383
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1797,
      "step": 5384
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.3546,
      "step": 5385
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1866,
      "step": 5386
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.295,
      "step": 5387
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1702,
      "step": 5388
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2896,
      "step": 5389
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.238,
      "step": 5390
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1775,
      "step": 5391
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1909,
      "step": 5392
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2298,
      "step": 5393
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2117,
      "step": 5394
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.3629,
      "step": 5395
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1741,
      "step": 5396
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.3099,
      "step": 5397
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2332,
      "step": 5398
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.3305,
      "step": 5399
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2128,
      "step": 5400
    },
    {
      "epoch": 1.12,
      "eval_loss": 0.3166283071041107,
      "eval_runtime": 159.3641,
      "eval_samples_per_second": 6.275,
      "eval_steps_per_second": 3.137,
      "step": 5400
    },
    {
      "epoch": 1.12,
      "mmlu_eval_accuracy": 0.5607196319490239,
      "mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.625,
      "mmlu_eval_accuracy_business_ethics": 0.7272727272727273,
      "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": 0.36363636363636365,
      "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": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
      "mmlu_eval_accuracy_econometrics": 0.3333333333333333,
      "mmlu_eval_accuracy_electrical_engineering": 0.25,
      "mmlu_eval_accuracy_elementary_mathematics": 0.5121951219512195,
      "mmlu_eval_accuracy_formal_logic": 0.07142857142857142,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.46875,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5116279069767442,
      "mmlu_eval_accuracy_high_school_mathematics": 0.13793103448275862,
      "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": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_world_history": 0.6153846153846154,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
      "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": 0.2727272727272727,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.33,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.5428571428571428,
      "mmlu_eval_accuracy_professional_accounting": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_law": 0.35294117647058826,
      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
      "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": 0.8636363636363636,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.772047083698427,
      "step": 5400
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.344,
      "step": 5401
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2547,
      "step": 5402
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.3143,
      "step": 5403
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1904,
      "step": 5404
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2698,
      "step": 5405
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2097,
      "step": 5406
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1936,
      "step": 5407
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2274,
      "step": 5408
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2491,
      "step": 5409
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2184,
      "step": 5410
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.3649,
      "step": 5411
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1805,
      "step": 5412
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1398,
      "step": 5413
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2113,
      "step": 5414
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1864,
      "step": 5415
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1969,
      "step": 5416
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1791,
      "step": 5417
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.1863,
      "step": 5418
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.179,
      "step": 5419
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2994,
      "step": 5420
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2433,
      "step": 5421
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.219,
      "step": 5422
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.4042,
      "step": 5423
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2076,
      "step": 5424
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2576,
      "step": 5425
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.242,
      "step": 5426
    },
    {
      "epoch": 1.12,
      "learning_rate": 0.0002,
      "loss": 0.2457,
      "step": 5427
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2523,
      "step": 5428
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.1858,
      "step": 5429
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.1906,
      "step": 5430
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2362,
      "step": 5431
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.1876,
      "step": 5432
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2906,
      "step": 5433
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.223,
      "step": 5434
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2122,
      "step": 5435
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2093,
      "step": 5436
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.1951,
      "step": 5437
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2405,
      "step": 5438
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.3071,
      "step": 5439
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.1487,
      "step": 5440
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2389,
      "step": 5441
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.1996,
      "step": 5442
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2021,
      "step": 5443
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.198,
      "step": 5444
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.1786,
      "step": 5445
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2708,
      "step": 5446
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.1425,
      "step": 5447
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2331,
      "step": 5448
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2323,
      "step": 5449
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2722,
      "step": 5450
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.1644,
      "step": 5451
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2271,
      "step": 5452
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2305,
      "step": 5453
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2235,
      "step": 5454
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2722,
      "step": 5455
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.188,
      "step": 5456
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.3754,
      "step": 5457
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.238,
      "step": 5458
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2016,
      "step": 5459
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2376,
      "step": 5460
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2478,
      "step": 5461
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2497,
      "step": 5462
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2618,
      "step": 5463
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2011,
      "step": 5464
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2325,
      "step": 5465
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.3791,
      "step": 5466
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2533,
      "step": 5467
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2242,
      "step": 5468
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2516,
      "step": 5469
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2511,
      "step": 5470
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.3417,
      "step": 5471
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2853,
      "step": 5472
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.287,
      "step": 5473
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2177,
      "step": 5474
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.1574,
      "step": 5475
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.0002,
      "loss": 0.2967,
      "step": 5476
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.312,
      "step": 5477
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2814,
      "step": 5478
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2277,
      "step": 5479
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.216,
      "step": 5480
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2853,
      "step": 5481
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1748,
      "step": 5482
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1974,
      "step": 5483
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2743,
      "step": 5484
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2689,
      "step": 5485
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1899,
      "step": 5486
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2798,
      "step": 5487
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1719,
      "step": 5488
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1809,
      "step": 5489
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1669,
      "step": 5490
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2382,
      "step": 5491
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.3688,
      "step": 5492
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.298,
      "step": 5493
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2611,
      "step": 5494
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2214,
      "step": 5495
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1525,
      "step": 5496
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.4067,
      "step": 5497
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1462,
      "step": 5498
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.284,
      "step": 5499
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2361,
      "step": 5500
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1745,
      "step": 5501
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.3118,
      "step": 5502
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.217,
      "step": 5503
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2757,
      "step": 5504
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2818,
      "step": 5505
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.3711,
      "step": 5506
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2055,
      "step": 5507
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2334,
      "step": 5508
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2112,
      "step": 5509
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1847,
      "step": 5510
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.3198,
      "step": 5511
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1976,
      "step": 5512
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.276,
      "step": 5513
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2622,
      "step": 5514
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2938,
      "step": 5515
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2826,
      "step": 5516
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2846,
      "step": 5517
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1993,
      "step": 5518
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2047,
      "step": 5519
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2857,
      "step": 5520
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2898,
      "step": 5521
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.1841,
      "step": 5522
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2084,
      "step": 5523
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0002,
      "loss": 0.2712,
      "step": 5524
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.1535,
      "step": 5525
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2093,
      "step": 5526
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.169,
      "step": 5527
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2291,
      "step": 5528
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.1444,
      "step": 5529
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.3931,
      "step": 5530
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2167,
      "step": 5531
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2353,
      "step": 5532
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.1488,
      "step": 5533
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.218,
      "step": 5534
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2168,
      "step": 5535
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2186,
      "step": 5536
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2313,
      "step": 5537
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.1696,
      "step": 5538
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2345,
      "step": 5539
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2203,
      "step": 5540
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.1596,
      "step": 5541
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.1688,
      "step": 5542
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2374,
      "step": 5543
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.3098,
      "step": 5544
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.3205,
      "step": 5545
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.3588,
      "step": 5546
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.192,
      "step": 5547
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2943,
      "step": 5548
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2097,
      "step": 5549
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2724,
      "step": 5550
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2905,
      "step": 5551
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.214,
      "step": 5552
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.208,
      "step": 5553
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.1769,
      "step": 5554
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2218,
      "step": 5555
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2663,
      "step": 5556
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2103,
      "step": 5557
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2077,
      "step": 5558
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2397,
      "step": 5559
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.339,
      "step": 5560
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2806,
      "step": 5561
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.1604,
      "step": 5562
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2347,
      "step": 5563
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2651,
      "step": 5564
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2311,
      "step": 5565
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2343,
      "step": 5566
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.1726,
      "step": 5567
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2668,
      "step": 5568
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.1979,
      "step": 5569
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.2206,
      "step": 5570
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.3262,
      "step": 5571
    },
    {
      "epoch": 1.15,
      "learning_rate": 0.0002,
      "loss": 0.1854,
      "step": 5572
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2688,
      "step": 5573
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2746,
      "step": 5574
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2099,
      "step": 5575
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2745,
      "step": 5576
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2051,
      "step": 5577
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2109,
      "step": 5578
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2,
      "step": 5579
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.3886,
      "step": 5580
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2755,
      "step": 5581
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2552,
      "step": 5582
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2406,
      "step": 5583
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2994,
      "step": 5584
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2054,
      "step": 5585
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.1981,
      "step": 5586
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2842,
      "step": 5587
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2645,
      "step": 5588
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2624,
      "step": 5589
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2544,
      "step": 5590
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2327,
      "step": 5591
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2528,
      "step": 5592
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.269,
      "step": 5593
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2545,
      "step": 5594
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.1767,
      "step": 5595
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2606,
      "step": 5596
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.3382,
      "step": 5597
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2215,
      "step": 5598
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2023,
      "step": 5599
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.0002,
      "loss": 0.2033,
      "step": 5600
    },
    {
      "epoch": 1.16,
      "eval_loss": 0.31649383902549744,
      "eval_runtime": 159.241,
      "eval_samples_per_second": 6.28,
      "eval_steps_per_second": 3.14,
      "step": 5600
    },
    {
      "epoch": 1.16,
      "mmlu_eval_accuracy": 0.5597542833051398,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.7272727272727273,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.5,
      "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": 0.5454545454545454,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.3333333333333333,
      "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": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5581395348837209,
      "mmlu_eval_accuracy_high_school_mathematics": 0.10344827586206896,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.3333333333333333,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.8,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7209302325581395,
      "mmlu_eval_accuracy_moral_disputes": 0.5,
      "mmlu_eval_accuracy_moral_scenarios": 0.32,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7647058823529411,
      "mmlu_eval_accuracy_prehistory": 0.6571428571428571,
      "mmlu_eval_accuracy_professional_accounting": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_law": 0.3764705882352941,
      "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": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.6111111111111112,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.6051943488127258,
      "step": 5600
    }
  ],
  "logging_steps": 1,
  "max_steps": 14472,
  "num_train_epochs": 3,
  "save_steps": 200,
  "total_flos": 1.3930480821556347e+18,
  "trial_name": null,
  "trial_params": null
}