Upload evals/e001_gpt2_124m_2p5B_s1339_dev.json with huggingface_hub
Browse files
evals/e001_gpt2_124m_2p5B_s1339_dev.json
ADDED
|
@@ -0,0 +1,107 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": "ckpt:runs/e001_gpt2_124m_2p5B_s1339/ckpt_004768.pt",
|
| 3 |
+
"split": "dev",
|
| 4 |
+
"tasks": {
|
| 5 |
+
"hellaswag": {
|
| 6 |
+
"n": 10042,
|
| 7 |
+
"acc": 0.2591117307309301,
|
| 8 |
+
"acc_ci95": [
|
| 9 |
+
0.25054521011750647,
|
| 10 |
+
0.26767576180043817
|
| 11 |
+
],
|
| 12 |
+
"acc_norm": 0.2746464847639912,
|
| 13 |
+
"acc_norm_ci95": [
|
| 14 |
+
0.2659803823939454,
|
| 15 |
+
0.28340967934674366
|
| 16 |
+
],
|
| 17 |
+
"gold_logprob_per_tok": -3.3829091590000875,
|
| 18 |
+
"gold_logprob_per_tok_ci95": [
|
| 19 |
+
-3.396688752443557,
|
| 20 |
+
-3.3707014061483775
|
| 21 |
+
],
|
| 22 |
+
"seconds": 31.6
|
| 23 |
+
},
|
| 24 |
+
"piqa": {
|
| 25 |
+
"n": 5000,
|
| 26 |
+
"acc": 0.6104,
|
| 27 |
+
"acc_ci95": [
|
| 28 |
+
0.5964,
|
| 29 |
+
0.6236050000000001
|
| 30 |
+
],
|
| 31 |
+
"acc_norm": 0.6106,
|
| 32 |
+
"acc_norm_ci95": [
|
| 33 |
+
0.596,
|
| 34 |
+
0.623805
|
| 35 |
+
],
|
| 36 |
+
"gold_logprob_per_tok": -4.168441109617508,
|
| 37 |
+
"gold_logprob_per_tok_ci95": [
|
| 38 |
+
-4.2088746506919055,
|
| 39 |
+
-4.131190399141974
|
| 40 |
+
],
|
| 41 |
+
"seconds": 6.8
|
| 42 |
+
},
|
| 43 |
+
"winogrande": {
|
| 44 |
+
"n": 5000,
|
| 45 |
+
"acc": 0.5182,
|
| 46 |
+
"acc_ci95": [
|
| 47 |
+
0.5054,
|
| 48 |
+
0.5322
|
| 49 |
+
],
|
| 50 |
+
"gold_logprob_per_tok": -3.7546719629464116,
|
| 51 |
+
"gold_logprob_per_tok_ci95": [
|
| 52 |
+
-3.7869353522446776,
|
| 53 |
+
-3.721578416062134
|
| 54 |
+
],
|
| 55 |
+
"seconds": 5.5
|
| 56 |
+
},
|
| 57 |
+
"arc_easy": {
|
| 58 |
+
"n": 2821,
|
| 59 |
+
"acc": 0.3775257001063453,
|
| 60 |
+
"acc_ci95": [
|
| 61 |
+
0.3598014888337469,
|
| 62 |
+
0.394549805033676
|
| 63 |
+
],
|
| 64 |
+
"acc_norm": 0.3271889400921659,
|
| 65 |
+
"acc_norm_ci95": [
|
| 66 |
+
0.30911024459411557,
|
| 67 |
+
0.34420418291386035
|
| 68 |
+
],
|
| 69 |
+
"gold_logprob_per_tok": -6.815205924929706,
|
| 70 |
+
"gold_logprob_per_tok_ci95": [
|
| 71 |
+
-6.926089525483299,
|
| 72 |
+
-6.703246408113291
|
| 73 |
+
],
|
| 74 |
+
"seconds": 7.0
|
| 75 |
+
},
|
| 76 |
+
"lambada": {
|
| 77 |
+
"n": 4869,
|
| 78 |
+
"acc": 0.16327788046826863,
|
| 79 |
+
"acc_ci95": [
|
| 80 |
+
0.15280345040049292,
|
| 81 |
+
0.17334154857260217
|
| 82 |
+
],
|
| 83 |
+
"ppl": 401.41456894665976,
|
| 84 |
+
"ppl_ci95": [
|
| 85 |
+
364.86376594916567,
|
| 86 |
+
440.85784238766615
|
| 87 |
+
],
|
| 88 |
+
"seconds": 12.7
|
| 89 |
+
},
|
| 90 |
+
"blimp": {
|
| 91 |
+
"n": 33500,
|
| 92 |
+
"acc": 0.7848059701492537,
|
| 93 |
+
"acc_ci95": [
|
| 94 |
+
0.7804134328358209,
|
| 95 |
+
0.788955223880597
|
| 96 |
+
],
|
| 97 |
+
"gold_logprob_per_tok": -6.459481815274967,
|
| 98 |
+
"gold_logprob_per_tok_ci95": [
|
| 99 |
+
-6.469976796309878,
|
| 100 |
+
-6.449362933883339
|
| 101 |
+
],
|
| 102 |
+
"seconds": 93.9
|
| 103 |
+
}
|
| 104 |
+
},
|
| 105 |
+
"suite_version": "evals-v0-dev",
|
| 106 |
+
"limit": null
|
| 107 |
+
}
|