diff --git "a/checkpoint-1000/trainer_state.json" "b/checkpoint-1000/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/checkpoint-1000/trainer_state.json"
@@ -0,0 +1,6106 @@
+{
+  "best_metric": 6.766034126281738,
+  "best_model_checkpoint": "./output_v2/34bCodellama_CodeLlama-34b-Python-hf_unnatural-instructions_standardized/checkpoint-1000",
+  "epoch": 0.007638835841417768,
+  "global_step": 1000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.0808,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.8773,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.1965,
+      "step": 3
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.118,
+      "step": 4
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.1773,
+      "step": 5
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.1165,
+      "step": 6
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.2666,
+      "step": 7
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.3704,
+      "step": 8
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.9976,
+      "step": 9
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.985,
+      "step": 10
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.0541,
+      "step": 11
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.6228,
+      "step": 12
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.3651,
+      "step": 13
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.0867,
+      "step": 14
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.4422,
+      "step": 15
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.7759,
+      "step": 16
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.1446,
+      "step": 17
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.0007,
+      "step": 18
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.0894,
+      "step": 19
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2424,
+      "step": 20
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.1343,
+      "step": 21
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.5354,
+      "step": 22
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.1887,
+      "step": 23
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.6652,
+      "step": 24
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.964,
+      "step": 25
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.1872,
+      "step": 26
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.4722,
+      "step": 27
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.1462,
+      "step": 28
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.0485,
+      "step": 29
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.148,
+      "step": 30
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.7274,
+      "step": 31
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.6689,
+      "step": 32
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.3384,
+      "step": 33
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.5354,
+      "step": 34
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.1976,
+      "step": 35
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.8593,
+      "step": 36
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.9302,
+      "step": 37
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.5968,
+      "step": 38
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.3169,
+      "step": 39
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.1793,
+      "step": 40
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.8457,
+      "step": 41
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.5177,
+      "step": 42
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.003,
+      "step": 43
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.9928,
+      "step": 44
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 15.2574,
+      "step": 45
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 0.3915,
+      "step": 46
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.4105,
+      "step": 47
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.1184,
+      "step": 48
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.72,
+      "step": 49
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.9628,
+      "step": 50
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.2372,
+      "step": 51
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3733,
+      "step": 52
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.8936,
+      "step": 53
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.5353,
+      "step": 54
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.0754,
+      "step": 55
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.6685,
+      "step": 56
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.8984,
+      "step": 57
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.2265,
+      "step": 58
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.7696,
+      "step": 59
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.7349,
+      "step": 60
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.0221,
+      "step": 61
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 15.1901,
+      "step": 62
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.387,
+      "step": 63
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.7323,
+      "step": 64
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.2077,
+      "step": 65
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.3155,
+      "step": 66
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.1656,
+      "step": 67
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 13.0828,
+      "step": 68
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.5295,
+      "step": 69
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.4575,
+      "step": 70
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 14.7654,
+      "step": 71
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.6263,
+      "step": 72
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 24.8238,
+      "step": 73
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 15.0654,
+      "step": 74
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 28.1046,
+      "step": 75
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 14.3232,
+      "step": 76
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 22.9712,
+      "step": 77
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 18.8529,
+      "step": 78
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 15.8356,
+      "step": 79
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 16.472,
+      "step": 80
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 12.2369,
+      "step": 81
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 14.0731,
+      "step": 82
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.8853,
+      "step": 83
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.5438,
+      "step": 84
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.2665,
+      "step": 85
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 11.5484,
+      "step": 86
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.7546,
+      "step": 87
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.4309,
+      "step": 88
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 11.5593,
+      "step": 89
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.3822,
+      "step": 90
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.6315,
+      "step": 91
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.6116,
+      "step": 92
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.2288,
+      "step": 93
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.0053,
+      "step": 94
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 12.359,
+      "step": 95
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9235,
+      "step": 96
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 31.9845,
+      "step": 97
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 11.1385,
+      "step": 98
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.6161,
+      "step": 99
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.8096,
+      "step": 100
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.9918,
+      "step": 101
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.344,
+      "step": 102
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.1607,
+      "step": 103
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.4834,
+      "step": 104
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.704,
+      "step": 105
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.1238,
+      "step": 106
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.8066,
+      "step": 107
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9656,
+      "step": 108
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.1979,
+      "step": 109
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2294,
+      "step": 110
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.066,
+      "step": 111
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7914,
+      "step": 112
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.7344,
+      "step": 113
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.6703,
+      "step": 114
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.8817,
+      "step": 115
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 11.7733,
+      "step": 116
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.469,
+      "step": 117
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.1304,
+      "step": 118
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.871,
+      "step": 119
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.5353,
+      "step": 120
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.9055,
+      "step": 121
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.6142,
+      "step": 122
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.0201,
+      "step": 123
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.3805,
+      "step": 124
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.6825,
+      "step": 125
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.7166,
+      "step": 126
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.7747,
+      "step": 127
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.7695,
+      "step": 128
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.7291,
+      "step": 129
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.1296,
+      "step": 130
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.5374,
+      "step": 131
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.1854,
+      "step": 132
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.434,
+      "step": 133
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.438,
+      "step": 134
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3027,
+      "step": 135
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.382,
+      "step": 136
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.9277,
+      "step": 137
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.223,
+      "step": 138
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.3042,
+      "step": 139
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.6361,
+      "step": 140
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.3547,
+      "step": 141
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.7181,
+      "step": 142
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.7528,
+      "step": 143
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 11.4316,
+      "step": 144
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.2219,
+      "step": 145
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.7788,
+      "step": 146
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.2749,
+      "step": 147
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.2397,
+      "step": 148
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.6243,
+      "step": 149
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.145,
+      "step": 150
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.7951,
+      "step": 151
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.1862,
+      "step": 152
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.1305,
+      "step": 153
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.5766,
+      "step": 154
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9232,
+      "step": 155
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.9936,
+      "step": 156
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.9692,
+      "step": 157
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.2772,
+      "step": 158
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.302,
+      "step": 159
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9931,
+      "step": 160
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.9675,
+      "step": 161
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.8536,
+      "step": 162
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.6589,
+      "step": 163
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.932,
+      "step": 164
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.0301,
+      "step": 165
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4861,
+      "step": 166
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.1354,
+      "step": 167
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0717,
+      "step": 168
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.9346,
+      "step": 169
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9373,
+      "step": 170
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.8777,
+      "step": 171
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.4193,
+      "step": 172
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.6831,
+      "step": 173
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4175,
+      "step": 174
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.3629,
+      "step": 175
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.118,
+      "step": 176
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.633,
+      "step": 177
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.8355,
+      "step": 178
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.4522,
+      "step": 179
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.9272,
+      "step": 180
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.4631,
+      "step": 181
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.2987,
+      "step": 182
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.1183,
+      "step": 183
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.9976,
+      "step": 184
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.0668,
+      "step": 185
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.6291,
+      "step": 186
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.5937,
+      "step": 187
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.7382,
+      "step": 188
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.7677,
+      "step": 189
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.0293,
+      "step": 190
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.6407,
+      "step": 191
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.9508,
+      "step": 192
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.5053,
+      "step": 193
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.5718,
+      "step": 194
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.5211,
+      "step": 195
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9557,
+      "step": 196
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.1609,
+      "step": 197
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.8505,
+      "step": 198
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.8278,
+      "step": 199
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.8447,
+      "step": 200
+    },
+    {
+      "epoch": 0.0,
+      "eval_loss": 7.883856773376465,
+      "eval_runtime": 22.4254,
+      "eval_samples_per_second": 2.23,
+      "eval_steps_per_second": 1.115,
+      "step": 200
+    },
+    {
+      "epoch": 0.0,
+      "mmlu_eval_accuracy": 0.2525477994227994,
+      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
+      "mmlu_eval_accuracy_anatomy": 0.07142857142857142,
+      "mmlu_eval_accuracy_astronomy": 0.3125,
+      "mmlu_eval_accuracy_business_ethics": 0.4444444444444444,
+      "mmlu_loss": 4.629522514343262,
+      "step": 200
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.3249,
+      "step": 201
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.352,
+      "step": 202
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2984,
+      "step": 203
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 11.2734,
+      "step": 204
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.1,
+      "step": 205
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.448,
+      "step": 206
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.2387,
+      "step": 207
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.861,
+      "step": 208
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.603,
+      "step": 209
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.29,
+      "step": 210
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2105,
+      "step": 211
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.1949,
+      "step": 212
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.0538,
+      "step": 213
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0343,
+      "step": 214
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.7794,
+      "step": 215
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.5532,
+      "step": 216
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2676,
+      "step": 217
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.566,
+      "step": 218
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.0432,
+      "step": 219
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9391,
+      "step": 220
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.724,
+      "step": 221
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.229,
+      "step": 222
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.3462,
+      "step": 223
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0752,
+      "step": 224
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.1966,
+      "step": 225
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7279,
+      "step": 226
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.8484,
+      "step": 227
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7291,
+      "step": 228
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.2665,
+      "step": 229
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.3551,
+      "step": 230
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.7338,
+      "step": 231
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.8407,
+      "step": 232
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.3581,
+      "step": 233
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.441,
+      "step": 234
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.0788,
+      "step": 235
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.8404,
+      "step": 236
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.4314,
+      "step": 237
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.8426,
+      "step": 238
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.0205,
+      "step": 239
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4162,
+      "step": 240
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.7515,
+      "step": 241
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.1442,
+      "step": 242
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.5868,
+      "step": 243
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.6514,
+      "step": 244
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2683,
+      "step": 245
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.31,
+      "step": 246
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0161,
+      "step": 247
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.484,
+      "step": 248
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.9726,
+      "step": 249
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.0926,
+      "step": 250
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.5279,
+      "step": 251
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.0017,
+      "step": 252
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.5684,
+      "step": 253
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.3875,
+      "step": 254
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.9489,
+      "step": 255
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.8948,
+      "step": 256
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.0856,
+      "step": 257
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.599,
+      "step": 258
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.1575,
+      "step": 259
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.3701,
+      "step": 260
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.464,
+      "step": 261
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9193,
+      "step": 262
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.5679,
+      "step": 263
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9424,
+      "step": 264
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.6689,
+      "step": 265
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.6475,
+      "step": 266
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.4311,
+      "step": 267
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.7426,
+      "step": 268
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.5191,
+      "step": 269
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.3059,
+      "step": 270
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.0142,
+      "step": 271
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.4509,
+      "step": 272
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.0831,
+      "step": 273
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.6977,
+      "step": 274
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.4236,
+      "step": 275
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2129,
+      "step": 276
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.1394,
+      "step": 277
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.685,
+      "step": 278
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.0275,
+      "step": 279
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 11.3215,
+      "step": 280
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.6542,
+      "step": 281
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.7614,
+      "step": 282
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.2996,
+      "step": 283
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.6275,
+      "step": 284
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.8736,
+      "step": 285
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.4667,
+      "step": 286
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.8486,
+      "step": 287
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2125,
+      "step": 288
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.4523,
+      "step": 289
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.551,
+      "step": 290
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 11.7158,
+      "step": 291
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.5092,
+      "step": 292
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9169,
+      "step": 293
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.5333,
+      "step": 294
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9949,
+      "step": 295
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.7189,
+      "step": 296
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.2366,
+      "step": 297
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.4745,
+      "step": 298
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.2439,
+      "step": 299
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4176,
+      "step": 300
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.9365,
+      "step": 301
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.5309,
+      "step": 302
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2201,
+      "step": 303
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.0312,
+      "step": 304
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.4173,
+      "step": 305
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.4856,
+      "step": 306
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.5041,
+      "step": 307
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.3597,
+      "step": 308
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.8395,
+      "step": 309
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.0776,
+      "step": 310
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.7566,
+      "step": 311
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.9767,
+      "step": 312
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.3804,
+      "step": 313
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 11.5327,
+      "step": 314
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.5293,
+      "step": 315
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4531,
+      "step": 316
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.3961,
+      "step": 317
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.5669,
+      "step": 318
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.8559,
+      "step": 319
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.117,
+      "step": 320
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.4279,
+      "step": 321
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.7977,
+      "step": 322
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.955,
+      "step": 323
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0164,
+      "step": 324
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 12.0495,
+      "step": 325
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.2768,
+      "step": 326
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.3162,
+      "step": 327
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.88,
+      "step": 328
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.2157,
+      "step": 329
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.8427,
+      "step": 330
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.9729,
+      "step": 331
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.1779,
+      "step": 332
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.1302,
+      "step": 333
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7705,
+      "step": 334
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.523,
+      "step": 335
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9375,
+      "step": 336
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.1409,
+      "step": 337
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.633,
+      "step": 338
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.6481,
+      "step": 339
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.933,
+      "step": 340
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.9179,
+      "step": 341
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.9332,
+      "step": 342
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.6553,
+      "step": 343
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.7412,
+      "step": 344
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.849,
+      "step": 345
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.7321,
+      "step": 346
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9717,
+      "step": 347
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.3465,
+      "step": 348
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4535,
+      "step": 349
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.2376,
+      "step": 350
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9025,
+      "step": 351
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.916,
+      "step": 352
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.3785,
+      "step": 353
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0576,
+      "step": 354
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.5081,
+      "step": 355
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.1303,
+      "step": 356
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.3854,
+      "step": 357
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 11.5553,
+      "step": 358
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9627,
+      "step": 359
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.402,
+      "step": 360
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3484,
+      "step": 361
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.5428,
+      "step": 362
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9128,
+      "step": 363
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3934,
+      "step": 364
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.4812,
+      "step": 365
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.5395,
+      "step": 366
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.6304,
+      "step": 367
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.5626,
+      "step": 368
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.5693,
+      "step": 369
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.3458,
+      "step": 370
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.6254,
+      "step": 371
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.8706,
+      "step": 372
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.6076,
+      "step": 373
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.2912,
+      "step": 374
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.3326,
+      "step": 375
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.3735,
+      "step": 376
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.4916,
+      "step": 377
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.5553,
+      "step": 378
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.6241,
+      "step": 379
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.6106,
+      "step": 380
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.266,
+      "step": 381
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7738,
+      "step": 382
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.4988,
+      "step": 383
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2968,
+      "step": 384
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.8512,
+      "step": 385
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0341,
+      "step": 386
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.898,
+      "step": 387
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.23,
+      "step": 388
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.9608,
+      "step": 389
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.3679,
+      "step": 390
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.7074,
+      "step": 391
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9903,
+      "step": 392
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.5845,
+      "step": 393
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.6493,
+      "step": 394
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7962,
+      "step": 395
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4865,
+      "step": 396
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.3418,
+      "step": 397
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.3942,
+      "step": 398
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.4715,
+      "step": 399
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.2073,
+      "step": 400
+    },
+    {
+      "epoch": 0.0,
+      "eval_loss": 7.106412410736084,
+      "eval_runtime": 22.5667,
+      "eval_samples_per_second": 2.216,
+      "eval_steps_per_second": 1.108,
+      "step": 400
+    },
+    {
+      "epoch": 0.0,
+      "mmlu_eval_accuracy": 0.2525477994227994,
+      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
+      "mmlu_eval_accuracy_anatomy": 0.07142857142857142,
+      "mmlu_eval_accuracy_astronomy": 0.3125,
+      "mmlu_eval_accuracy_business_ethics": 0.4444444444444444,
+      "mmlu_loss": 2.9128687667846678,
+      "step": 400
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.3984,
+      "step": 401
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.7983,
+      "step": 402
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.8589,
+      "step": 403
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.9884,
+      "step": 404
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.4427,
+      "step": 405
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0374,
+      "step": 406
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.7999,
+      "step": 407
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.2437,
+      "step": 408
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.6902,
+      "step": 409
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.81,
+      "step": 410
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.8979,
+      "step": 411
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0211,
+      "step": 412
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.3945,
+      "step": 413
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.5807,
+      "step": 414
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.1433,
+      "step": 415
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9466,
+      "step": 416
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.6276,
+      "step": 417
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.4945,
+      "step": 418
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.6215,
+      "step": 419
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.3919,
+      "step": 420
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.7915,
+      "step": 421
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.3284,
+      "step": 422
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.8723,
+      "step": 423
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.0149,
+      "step": 424
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.979,
+      "step": 425
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.9175,
+      "step": 426
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.4994,
+      "step": 427
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.9791,
+      "step": 428
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.1156,
+      "step": 429
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.5813,
+      "step": 430
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.1882,
+      "step": 431
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.9956,
+      "step": 432
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.6189,
+      "step": 433
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.9624,
+      "step": 434
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.5387,
+      "step": 435
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4605,
+      "step": 436
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.474,
+      "step": 437
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0497,
+      "step": 438
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.5705,
+      "step": 439
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.275,
+      "step": 440
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.9638,
+      "step": 441
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.4857,
+      "step": 442
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3067,
+      "step": 443
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.8152,
+      "step": 444
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.1668,
+      "step": 445
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.5293,
+      "step": 446
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3981,
+      "step": 447
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.4787,
+      "step": 448
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.5981,
+      "step": 449
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.3569,
+      "step": 450
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4088,
+      "step": 451
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.3677,
+      "step": 452
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.4686,
+      "step": 453
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.3552,
+      "step": 454
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.7931,
+      "step": 455
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.9285,
+      "step": 456
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.0554,
+      "step": 457
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.7277,
+      "step": 458
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2474,
+      "step": 459
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.9274,
+      "step": 460
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.2558,
+      "step": 461
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.7547,
+      "step": 462
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.1264,
+      "step": 463
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.2124,
+      "step": 464
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.8751,
+      "step": 465
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.7317,
+      "step": 466
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3697,
+      "step": 467
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0021,
+      "step": 468
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3761,
+      "step": 469
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2291,
+      "step": 470
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7968,
+      "step": 471
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.9454,
+      "step": 472
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.0194,
+      "step": 473
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.5048,
+      "step": 474
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.6837,
+      "step": 475
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.1066,
+      "step": 476
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.3501,
+      "step": 477
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.5071,
+      "step": 478
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.1086,
+      "step": 479
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.7269,
+      "step": 480
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.5419,
+      "step": 481
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.2974,
+      "step": 482
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.1433,
+      "step": 483
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0869,
+      "step": 484
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.032,
+      "step": 485
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.0946,
+      "step": 486
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.7162,
+      "step": 487
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.0406,
+      "step": 488
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.9048,
+      "step": 489
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.2231,
+      "step": 490
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.6524,
+      "step": 491
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.1151,
+      "step": 492
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.591,
+      "step": 493
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.1628,
+      "step": 494
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.0757,
+      "step": 495
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.3471,
+      "step": 496
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9385,
+      "step": 497
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9362,
+      "step": 498
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.2252,
+      "step": 499
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.359,
+      "step": 500
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0497,
+      "step": 501
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.0484,
+      "step": 502
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.5773,
+      "step": 503
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.39,
+      "step": 504
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.5923,
+      "step": 505
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2,
+      "step": 506
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.5536,
+      "step": 507
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.8958,
+      "step": 508
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7763,
+      "step": 509
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.2045,
+      "step": 510
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.4219,
+      "step": 511
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.6305,
+      "step": 512
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.4243,
+      "step": 513
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7842,
+      "step": 514
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.8769,
+      "step": 515
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.8903,
+      "step": 516
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.0489,
+      "step": 517
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.1314,
+      "step": 518
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.5973,
+      "step": 519
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.8022,
+      "step": 520
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3539,
+      "step": 521
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.222,
+      "step": 522
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.5403,
+      "step": 523
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.1323,
+      "step": 524
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7813,
+      "step": 525
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.4982,
+      "step": 526
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2426,
+      "step": 527
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.0142,
+      "step": 528
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.8996,
+      "step": 529
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.8671,
+      "step": 530
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.4139,
+      "step": 531
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9478,
+      "step": 532
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.7062,
+      "step": 533
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.0098,
+      "step": 534
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.9195,
+      "step": 535
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.0255,
+      "step": 536
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.6291,
+      "step": 537
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.3245,
+      "step": 538
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.6382,
+      "step": 539
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 1.8076,
+      "step": 540
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.6725,
+      "step": 541
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.0563,
+      "step": 542
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.6178,
+      "step": 543
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.7974,
+      "step": 544
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.7535,
+      "step": 545
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4948,
+      "step": 546
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.8941,
+      "step": 547
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.6496,
+      "step": 548
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.9084,
+      "step": 549
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.65,
+      "step": 550
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7673,
+      "step": 551
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 10.2221,
+      "step": 552
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.14,
+      "step": 553
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.6747,
+      "step": 554
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.8009,
+      "step": 555
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.7307,
+      "step": 556
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.0143,
+      "step": 557
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.8098,
+      "step": 558
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.026,
+      "step": 559
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.4572,
+      "step": 560
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.7913,
+      "step": 561
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9962,
+      "step": 562
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.767,
+      "step": 563
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.9497,
+      "step": 564
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.9626,
+      "step": 565
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2536,
+      "step": 566
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.0421,
+      "step": 567
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.8177,
+      "step": 568
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9241,
+      "step": 569
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.0162,
+      "step": 570
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.3368,
+      "step": 571
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7515,
+      "step": 572
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.6389,
+      "step": 573
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.662,
+      "step": 574
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.8097,
+      "step": 575
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9346,
+      "step": 576
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.3154,
+      "step": 577
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.7724,
+      "step": 578
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3685,
+      "step": 579
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.2775,
+      "step": 580
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.106,
+      "step": 581
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.4733,
+      "step": 582
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.2334,
+      "step": 583
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9478,
+      "step": 584
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.0013,
+      "step": 585
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.7242,
+      "step": 586
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.922,
+      "step": 587
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.1418,
+      "step": 588
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4472,
+      "step": 589
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.4785,
+      "step": 590
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.783,
+      "step": 591
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.0706,
+      "step": 592
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.4136,
+      "step": 593
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.5969,
+      "step": 594
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.5157,
+      "step": 595
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.5658,
+      "step": 596
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.4647,
+      "step": 597
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.2028,
+      "step": 598
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.6913,
+      "step": 599
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.7239,
+      "step": 600
+    },
+    {
+      "epoch": 0.0,
+      "eval_loss": 7.012163162231445,
+      "eval_runtime": 22.5807,
+      "eval_samples_per_second": 2.214,
+      "eval_steps_per_second": 1.107,
+      "step": 600
+    },
+    {
+      "epoch": 0.0,
+      "mmlu_eval_accuracy": 0.3260281385281385,
+      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
+      "mmlu_eval_accuracy_anatomy": 0.35714285714285715,
+      "mmlu_eval_accuracy_astronomy": 0.25,
+      "mmlu_eval_accuracy_business_ethics": 0.3333333333333333,
+      "mmlu_loss": 4.24488224029541,
+      "step": 600
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.5253,
+      "step": 601
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0392,
+      "step": 602
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.447,
+      "step": 603
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.9441,
+      "step": 604
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.1874,
+      "step": 605
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.7817,
+      "step": 606
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.0348,
+      "step": 607
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.5593,
+      "step": 608
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.9361,
+      "step": 609
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3534,
+      "step": 610
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.476,
+      "step": 611
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.0937,
+      "step": 612
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3027,
+      "step": 613
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.5586,
+      "step": 614
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3796,
+      "step": 615
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.676,
+      "step": 616
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.5321,
+      "step": 617
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.0059,
+      "step": 618
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.6139,
+      "step": 619
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.2391,
+      "step": 620
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.0636,
+      "step": 621
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.0895,
+      "step": 622
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.62,
+      "step": 623
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.0469,
+      "step": 624
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.2173,
+      "step": 625
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 3.9432,
+      "step": 626
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.3928,
+      "step": 627
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.0959,
+      "step": 628
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.1197,
+      "step": 629
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 2.4277,
+      "step": 630
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.418,
+      "step": 631
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.8687,
+      "step": 632
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.0156,
+      "step": 633
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.573,
+      "step": 634
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.112,
+      "step": 635
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.8954,
+      "step": 636
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.36,
+      "step": 637
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.924,
+      "step": 638
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.4625,
+      "step": 639
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.2023,
+      "step": 640
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.0685,
+      "step": 641
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.5304,
+      "step": 642
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4456,
+      "step": 643
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.7271,
+      "step": 644
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.6011,
+      "step": 645
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.895,
+      "step": 646
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 4.864,
+      "step": 647
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.3452,
+      "step": 648
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 5.8978,
+      "step": 649
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.2253,
+      "step": 650
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 9.2813,
+      "step": 651
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 7.7248,
+      "step": 652
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 8.4283,
+      "step": 653
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0004,
+      "loss": 6.4304,
+      "step": 654
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.3893,
+      "step": 655
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.1115,
+      "step": 656
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.5892,
+      "step": 657
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.6572,
+      "step": 658
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.925,
+      "step": 659
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.4431,
+      "step": 660
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.7711,
+      "step": 661
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.9439,
+      "step": 662
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.3781,
+      "step": 663
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.5573,
+      "step": 664
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 10.4476,
+      "step": 665
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.0057,
+      "step": 666
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.2702,
+      "step": 667
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.5717,
+      "step": 668
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.2242,
+      "step": 669
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.1,
+      "step": 670
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.0517,
+      "step": 671
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.6543,
+      "step": 672
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.1138,
+      "step": 673
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.461,
+      "step": 674
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.7094,
+      "step": 675
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.521,
+      "step": 676
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.7116,
+      "step": 677
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.6343,
+      "step": 678
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.3762,
+      "step": 679
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.3603,
+      "step": 680
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.7144,
+      "step": 681
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.4545,
+      "step": 682
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.8188,
+      "step": 683
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.7965,
+      "step": 684
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 2.4675,
+      "step": 685
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.0436,
+      "step": 686
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.1219,
+      "step": 687
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.4517,
+      "step": 688
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.8476,
+      "step": 689
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 1.9284,
+      "step": 690
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.7405,
+      "step": 691
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.7142,
+      "step": 692
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.3979,
+      "step": 693
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 1.3285,
+      "step": 694
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.3418,
+      "step": 695
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.4472,
+      "step": 696
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.7355,
+      "step": 697
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.7982,
+      "step": 698
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 2.4516,
+      "step": 699
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 2.2532,
+      "step": 700
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.9959,
+      "step": 701
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.0418,
+      "step": 702
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 10.7767,
+      "step": 703
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.774,
+      "step": 704
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.8912,
+      "step": 705
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.2178,
+      "step": 706
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.6197,
+      "step": 707
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.4755,
+      "step": 708
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.8276,
+      "step": 709
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.2925,
+      "step": 710
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.3887,
+      "step": 711
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.1465,
+      "step": 712
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.5806,
+      "step": 713
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.3063,
+      "step": 714
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.6066,
+      "step": 715
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.1536,
+      "step": 716
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.5582,
+      "step": 717
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.0353,
+      "step": 718
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.6415,
+      "step": 719
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.8291,
+      "step": 720
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 10.7575,
+      "step": 721
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.9141,
+      "step": 722
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.5217,
+      "step": 723
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.4549,
+      "step": 724
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.8112,
+      "step": 725
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.2729,
+      "step": 726
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.8515,
+      "step": 727
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.9712,
+      "step": 728
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.097,
+      "step": 729
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.0208,
+      "step": 730
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.1377,
+      "step": 731
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.4019,
+      "step": 732
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.9869,
+      "step": 733
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.2954,
+      "step": 734
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.4144,
+      "step": 735
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.8053,
+      "step": 736
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.8891,
+      "step": 737
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.812,
+      "step": 738
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.2657,
+      "step": 739
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.3747,
+      "step": 740
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.0364,
+      "step": 741
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.8845,
+      "step": 742
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.887,
+      "step": 743
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.0706,
+      "step": 744
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.6619,
+      "step": 745
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.2941,
+      "step": 746
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.9192,
+      "step": 747
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.9947,
+      "step": 748
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.6376,
+      "step": 749
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.0358,
+      "step": 750
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.4578,
+      "step": 751
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.7784,
+      "step": 752
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.632,
+      "step": 753
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.8649,
+      "step": 754
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.7951,
+      "step": 755
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.3841,
+      "step": 756
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.4558,
+      "step": 757
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.7638,
+      "step": 758
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.9413,
+      "step": 759
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.0916,
+      "step": 760
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.1351,
+      "step": 761
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.6078,
+      "step": 762
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.7982,
+      "step": 763
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.6132,
+      "step": 764
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.551,
+      "step": 765
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.3301,
+      "step": 766
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.4888,
+      "step": 767
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.1476,
+      "step": 768
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.4244,
+      "step": 769
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.6025,
+      "step": 770
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.102,
+      "step": 771
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.017,
+      "step": 772
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.4101,
+      "step": 773
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.1741,
+      "step": 774
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.1256,
+      "step": 775
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.5164,
+      "step": 776
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 2.6959,
+      "step": 777
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.7666,
+      "step": 778
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.4336,
+      "step": 779
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 11.8478,
+      "step": 780
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.8382,
+      "step": 781
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.1792,
+      "step": 782
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.4424,
+      "step": 783
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.345,
+      "step": 784
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.6887,
+      "step": 785
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.9867,
+      "step": 786
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.6152,
+      "step": 787
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.7283,
+      "step": 788
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.0157,
+      "step": 789
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.6044,
+      "step": 790
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.4132,
+      "step": 791
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.735,
+      "step": 792
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.3631,
+      "step": 793
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.2308,
+      "step": 794
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.2184,
+      "step": 795
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.4661,
+      "step": 796
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.9637,
+      "step": 797
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.4178,
+      "step": 798
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.5909,
+      "step": 799
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.1482,
+      "step": 800
+    },
+    {
+      "epoch": 0.01,
+      "eval_loss": 7.355834484100342,
+      "eval_runtime": 22.6252,
+      "eval_samples_per_second": 2.21,
+      "eval_steps_per_second": 1.105,
+      "step": 800
+    },
+    {
+      "epoch": 0.01,
+      "mmlu_eval_accuracy": 0.2525477994227994,
+      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
+      "mmlu_eval_accuracy_anatomy": 0.07142857142857142,
+      "mmlu_eval_accuracy_astronomy": 0.3125,
+      "mmlu_eval_accuracy_business_ethics": 0.4444444444444444,
+      "mmlu_loss": 5.191131496429444,
+      "step": 800
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 10.0427,
+      "step": 801
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.2669,
+      "step": 802
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.8026,
+      "step": 803
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.4949,
+      "step": 804
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.4491,
+      "step": 805
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.0383,
+      "step": 806
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.1213,
+      "step": 807
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.5158,
+      "step": 808
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.5648,
+      "step": 809
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.9932,
+      "step": 810
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.6441,
+      "step": 811
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.8661,
+      "step": 812
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.3609,
+      "step": 813
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.6828,
+      "step": 814
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.9693,
+      "step": 815
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.3733,
+      "step": 816
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.6286,
+      "step": 817
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.4349,
+      "step": 818
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.6706,
+      "step": 819
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.3089,
+      "step": 820
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.2394,
+      "step": 821
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.963,
+      "step": 822
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.6564,
+      "step": 823
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.997,
+      "step": 824
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 1.9261,
+      "step": 825
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.1421,
+      "step": 826
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.2335,
+      "step": 827
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.3432,
+      "step": 828
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.0154,
+      "step": 829
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.5135,
+      "step": 830
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.6226,
+      "step": 831
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.1672,
+      "step": 832
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.0853,
+      "step": 833
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.1213,
+      "step": 834
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.7815,
+      "step": 835
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.8916,
+      "step": 836
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.6464,
+      "step": 837
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.3307,
+      "step": 838
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.8165,
+      "step": 839
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 2.886,
+      "step": 840
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.4781,
+      "step": 841
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 2.8351,
+      "step": 842
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.358,
+      "step": 843
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.6501,
+      "step": 844
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.0864,
+      "step": 845
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.2922,
+      "step": 846
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.9847,
+      "step": 847
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.2558,
+      "step": 848
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.0195,
+      "step": 849
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.996,
+      "step": 850
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.5705,
+      "step": 851
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.4136,
+      "step": 852
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.6302,
+      "step": 853
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.8761,
+      "step": 854
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.4995,
+      "step": 855
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.4762,
+      "step": 856
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.5749,
+      "step": 857
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.0273,
+      "step": 858
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.8258,
+      "step": 859
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.1836,
+      "step": 860
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.5493,
+      "step": 861
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.1891,
+      "step": 862
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.7392,
+      "step": 863
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.1655,
+      "step": 864
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.5218,
+      "step": 865
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.3759,
+      "step": 866
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.2497,
+      "step": 867
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.5901,
+      "step": 868
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.0624,
+      "step": 869
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 10.2452,
+      "step": 870
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.5649,
+      "step": 871
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.0826,
+      "step": 872
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.2703,
+      "step": 873
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.9088,
+      "step": 874
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.3875,
+      "step": 875
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.2511,
+      "step": 876
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.4065,
+      "step": 877
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.175,
+      "step": 878
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.8358,
+      "step": 879
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.3208,
+      "step": 880
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.2049,
+      "step": 881
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.8251,
+      "step": 882
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.4262,
+      "step": 883
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.2227,
+      "step": 884
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.1062,
+      "step": 885
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.9417,
+      "step": 886
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.3315,
+      "step": 887
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.0012,
+      "step": 888
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.6386,
+      "step": 889
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.0377,
+      "step": 890
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.6707,
+      "step": 891
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.4955,
+      "step": 892
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.7343,
+      "step": 893
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.8305,
+      "step": 894
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.7016,
+      "step": 895
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.7149,
+      "step": 896
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.5649,
+      "step": 897
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.815,
+      "step": 898
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.6135,
+      "step": 899
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.8776,
+      "step": 900
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.7288,
+      "step": 901
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.8019,
+      "step": 902
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.0921,
+      "step": 903
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.824,
+      "step": 904
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.7151,
+      "step": 905
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.5054,
+      "step": 906
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.8095,
+      "step": 907
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.3218,
+      "step": 908
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.9993,
+      "step": 909
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.4433,
+      "step": 910
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.5863,
+      "step": 911
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.505,
+      "step": 912
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.9734,
+      "step": 913
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.1792,
+      "step": 914
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.4574,
+      "step": 915
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.2787,
+      "step": 916
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.8201,
+      "step": 917
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.2537,
+      "step": 918
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.1387,
+      "step": 919
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.7161,
+      "step": 920
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.2207,
+      "step": 921
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.7953,
+      "step": 922
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.9949,
+      "step": 923
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.9173,
+      "step": 924
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.7903,
+      "step": 925
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.4784,
+      "step": 926
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.2264,
+      "step": 927
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.566,
+      "step": 928
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.0686,
+      "step": 929
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 2.791,
+      "step": 930
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.8393,
+      "step": 931
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.4387,
+      "step": 932
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.2374,
+      "step": 933
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.9598,
+      "step": 934
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.1597,
+      "step": 935
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 9.0403,
+      "step": 936
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.3301,
+      "step": 937
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.072,
+      "step": 938
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.4788,
+      "step": 939
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.0656,
+      "step": 940
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.9647,
+      "step": 941
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.1168,
+      "step": 942
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.0293,
+      "step": 943
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.3622,
+      "step": 944
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.8957,
+      "step": 945
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.4,
+      "step": 946
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 2.6626,
+      "step": 947
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.8212,
+      "step": 948
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.8638,
+      "step": 949
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.6406,
+      "step": 950
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.7069,
+      "step": 951
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.1384,
+      "step": 952
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.612,
+      "step": 953
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.7201,
+      "step": 954
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.3532,
+      "step": 955
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.1266,
+      "step": 956
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.6192,
+      "step": 957
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.826,
+      "step": 958
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.9338,
+      "step": 959
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.4487,
+      "step": 960
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.872,
+      "step": 961
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.8601,
+      "step": 962
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.7401,
+      "step": 963
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.5412,
+      "step": 964
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.2501,
+      "step": 965
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.6837,
+      "step": 966
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.6494,
+      "step": 967
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.604,
+      "step": 968
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.837,
+      "step": 969
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.3957,
+      "step": 970
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.3281,
+      "step": 971
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.8264,
+      "step": 972
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.6554,
+      "step": 973
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.5768,
+      "step": 974
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.4187,
+      "step": 975
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.8479,
+      "step": 976
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.9849,
+      "step": 977
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.6471,
+      "step": 978
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.8041,
+      "step": 979
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.8876,
+      "step": 980
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.6423,
+      "step": 981
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.5329,
+      "step": 982
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.2801,
+      "step": 983
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.1699,
+      "step": 984
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.6469,
+      "step": 985
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.6766,
+      "step": 986
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.7538,
+      "step": 987
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.9606,
+      "step": 988
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 3.0713,
+      "step": 989
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 7.4965,
+      "step": 990
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.3408,
+      "step": 991
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.4007,
+      "step": 992
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 2.8921,
+      "step": 993
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.8681,
+      "step": 994
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 1.8867,
+      "step": 995
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.467,
+      "step": 996
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 8.7895,
+      "step": 997
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 6.0523,
+      "step": 998
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 4.4032,
+      "step": 999
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0004,
+      "loss": 5.7719,
+      "step": 1000
+    },
+    {
+      "epoch": 0.01,
+      "eval_loss": 6.766034126281738,
+      "eval_runtime": 22.4042,
+      "eval_samples_per_second": 2.232,
+      "eval_steps_per_second": 1.116,
+      "step": 1000
+    },
+    {
+      "epoch": 0.01,
+      "mmlu_eval_accuracy": 0.2525477994227994,
+      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
+      "mmlu_eval_accuracy_anatomy": 0.07142857142857142,
+      "mmlu_eval_accuracy_astronomy": 0.3125,
+      "mmlu_eval_accuracy_business_ethics": 0.4444444444444444,
+      "mmlu_loss": 4.338861379623413,
+      "step": 1000
+    }
+  ],
+  "max_steps": 30000,
+  "num_train_epochs": 1,
+  "total_flos": 1.6844425298116608e+16,
+  "trial_name": null,
+  "trial_params": null
+}