Upload evals/e001_gpt2_124m_2p5B_s1337_dev.json with huggingface_hub
Browse files
evals/e001_gpt2_124m_2p5B_s1337_dev.json
ADDED
|
@@ -0,0 +1,107 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": "ckpt:runs/e001_gpt2_124m_2p5B_s1337/ckpt_004768.pt",
|
| 3 |
+
"split": "dev",
|
| 4 |
+
"tasks": {
|
| 5 |
+
"hellaswag": {
|
| 6 |
+
"n": 10042,
|
| 7 |
+
"acc": 0.25941047600079664,
|
| 8 |
+
"acc_ci95": [
|
| 9 |
+
0.25124477195777734,
|
| 10 |
+
0.2677778331009759
|
| 11 |
+
],
|
| 12 |
+
"acc_norm": 0.276638119896435,
|
| 13 |
+
"acc_norm_ci95": [
|
| 14 |
+
0.26777285401314477,
|
| 15 |
+
0.28510256920932087
|
| 16 |
+
],
|
| 17 |
+
"gold_logprob_per_tok": -3.360447519031279,
|
| 18 |
+
"gold_logprob_per_tok_ci95": [
|
| 19 |
+
-3.3740750079927024,
|
| 20 |
+
-3.347677461746153
|
| 21 |
+
],
|
| 22 |
+
"seconds": 31.5
|
| 23 |
+
},
|
| 24 |
+
"piqa": {
|
| 25 |
+
"n": 5000,
|
| 26 |
+
"acc": 0.6084,
|
| 27 |
+
"acc_ci95": [
|
| 28 |
+
0.5958,
|
| 29 |
+
0.6224
|
| 30 |
+
],
|
| 31 |
+
"acc_norm": 0.617,
|
| 32 |
+
"acc_norm_ci95": [
|
| 33 |
+
0.6033900000000001,
|
| 34 |
+
0.62901
|
| 35 |
+
],
|
| 36 |
+
"gold_logprob_per_tok": -4.121109387697325,
|
| 37 |
+
"gold_logprob_per_tok_ci95": [
|
| 38 |
+
-4.162696102944079,
|
| 39 |
+
-4.084296691647292
|
| 40 |
+
],
|
| 41 |
+
"seconds": 8.0
|
| 42 |
+
},
|
| 43 |
+
"winogrande": {
|
| 44 |
+
"n": 5000,
|
| 45 |
+
"acc": 0.5122,
|
| 46 |
+
"acc_ci95": [
|
| 47 |
+
0.498195,
|
| 48 |
+
0.526
|
| 49 |
+
],
|
| 50 |
+
"gold_logprob_per_tok": -3.7478074397976706,
|
| 51 |
+
"gold_logprob_per_tok_ci95": [
|
| 52 |
+
-3.781759251461203,
|
| 53 |
+
-3.7135984887131808
|
| 54 |
+
],
|
| 55 |
+
"seconds": 6.8
|
| 56 |
+
},
|
| 57 |
+
"arc_easy": {
|
| 58 |
+
"n": 2821,
|
| 59 |
+
"acc": 0.38780574264445233,
|
| 60 |
+
"acc_ci95": [
|
| 61 |
+
0.3697181850407657,
|
| 62 |
+
0.40482984757178303
|
| 63 |
+
],
|
| 64 |
+
"acc_norm": 0.3424317617866005,
|
| 65 |
+
"acc_norm_ci95": [
|
| 66 |
+
0.32469868840836585,
|
| 67 |
+
0.36016483516483516
|
| 68 |
+
],
|
| 69 |
+
"gold_logprob_per_tok": -6.679924925094083,
|
| 70 |
+
"gold_logprob_per_tok_ci95": [
|
| 71 |
+
-6.7895326975915005,
|
| 72 |
+
-6.5670429001935275
|
| 73 |
+
],
|
| 74 |
+
"seconds": 8.4
|
| 75 |
+
},
|
| 76 |
+
"lambada": {
|
| 77 |
+
"n": 4869,
|
| 78 |
+
"acc": 0.16820702402957485,
|
| 79 |
+
"acc_ci95": [
|
| 80 |
+
0.15834360238241937,
|
| 81 |
+
0.17785993017046622
|
| 82 |
+
],
|
| 83 |
+
"ppl": 380.0502088197219,
|
| 84 |
+
"ppl_ci95": [
|
| 85 |
+
346.3778265564967,
|
| 86 |
+
417.0042814734404
|
| 87 |
+
],
|
| 88 |
+
"seconds": 17.1
|
| 89 |
+
},
|
| 90 |
+
"blimp": {
|
| 91 |
+
"n": 33500,
|
| 92 |
+
"acc": 0.7883582089552239,
|
| 93 |
+
"acc_ci95": [
|
| 94 |
+
0.784179104477612,
|
| 95 |
+
0.7931358208955224
|
| 96 |
+
],
|
| 97 |
+
"gold_logprob_per_tok": -6.490564379815169,
|
| 98 |
+
"gold_logprob_per_tok_ci95": [
|
| 99 |
+
-6.501086070166422,
|
| 100 |
+
-6.480123320572972
|
| 101 |
+
],
|
| 102 |
+
"seconds": 126.8
|
| 103 |
+
}
|
| 104 |
+
},
|
| 105 |
+
"suite_version": "evals-v0-dev",
|
| 106 |
+
"limit": null
|
| 107 |
+
}
|