Upload _paper_results/reasoning_rl_multiseed_summary.json with huggingface_hub
Browse files
_paper_results/reasoning_rl_multiseed_summary.json
ADDED
|
@@ -0,0 +1,111 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"t1": {
|
| 3 |
+
"per_seed": [
|
| 4 |
+
{
|
| 5 |
+
"seed": 2,
|
| 6 |
+
"score_path": "/workspace/dnathinker/runs/eval_reasoning_t1_v7r128_postRL_alpha1_s2_20260506_203542/score.json",
|
| 7 |
+
"tfg": 0.3964326508231607,
|
| 8 |
+
"n_cited": 20.48,
|
| 9 |
+
"n_grounded": 9.68,
|
| 10 |
+
"n_halluc": 10.58,
|
| 11 |
+
"reasoning_tags_rate": 0.6
|
| 12 |
+
},
|
| 13 |
+
{
|
| 14 |
+
"seed": 3,
|
| 15 |
+
"score_path": "/workspace/dnathinker/runs/eval_reasoning_t1_v7r128_postRL_alpha1_s3_20260506_234027/score.json",
|
| 16 |
+
"tfg": 0.3996367124792261,
|
| 17 |
+
"n_cited": 22.84,
|
| 18 |
+
"n_grounded": 13.74,
|
| 19 |
+
"n_halluc": 8.76,
|
| 20 |
+
"reasoning_tags_rate": 0.2
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"seed": 42,
|
| 24 |
+
"score_path": "/workspace/dnathinker/runs/eval_reasoning_t1_v7r128_postRLext_best_20260506_053915/score.json",
|
| 25 |
+
"tfg": 0.43835448568269475,
|
| 26 |
+
"n_cited": 27.46,
|
| 27 |
+
"n_grounded": 14.58,
|
| 28 |
+
"n_halluc": 12.36,
|
| 29 |
+
"reasoning_tags_rate": 0.12
|
| 30 |
+
}
|
| 31 |
+
],
|
| 32 |
+
"tfg_stats": {
|
| 33 |
+
"n": 3,
|
| 34 |
+
"mean": 0.4114746163283605,
|
| 35 |
+
"std": 0.023333710274054672,
|
| 36 |
+
"min": 0.3964326508231607,
|
| 37 |
+
"max": 0.43835448568269475,
|
| 38 |
+
"values": [
|
| 39 |
+
0.3964326508231607,
|
| 40 |
+
0.3996367124792261,
|
| 41 |
+
0.43835448568269475
|
| 42 |
+
]
|
| 43 |
+
}
|
| 44 |
+
},
|
| 45 |
+
"t2": {
|
| 46 |
+
"per_seed": [
|
| 47 |
+
{
|
| 48 |
+
"seed": 42,
|
| 49 |
+
"score_path": "/workspace/dnathinker/runs/eval_reasoning_t2_v7r128_postRL_alpha1_20260506_004858/score.json",
|
| 50 |
+
"tfg": 0.3650301689387592,
|
| 51 |
+
"n_cited": 20.58,
|
| 52 |
+
"n_grounded": 9.84,
|
| 53 |
+
"n_halluc": 10.08,
|
| 54 |
+
"reasoning_tags_rate": 0.44
|
| 55 |
+
}
|
| 56 |
+
],
|
| 57 |
+
"tfg_stats": {
|
| 58 |
+
"n": 1,
|
| 59 |
+
"mean": 0.3650301689387592,
|
| 60 |
+
"std": 0.0,
|
| 61 |
+
"min": 0.3650301689387592,
|
| 62 |
+
"max": 0.3650301689387592,
|
| 63 |
+
"values": [
|
| 64 |
+
0.3650301689387592
|
| 65 |
+
]
|
| 66 |
+
}
|
| 67 |
+
},
|
| 68 |
+
"t3": {
|
| 69 |
+
"per_seed": [
|
| 70 |
+
{
|
| 71 |
+
"seed": 2,
|
| 72 |
+
"score_path": "/workspace/dnathinker/runs/eval_reasoning_t3_v7r128_postRL_alpha1_s2_par_20260506_221253/score.json",
|
| 73 |
+
"tfg": 0.23970945554855788,
|
| 74 |
+
"n_cited": 14.38,
|
| 75 |
+
"n_grounded": 6.46,
|
| 76 |
+
"n_halluc": 7.52,
|
| 77 |
+
"reasoning_tags_rate": 0.64
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"seed": 3,
|
| 81 |
+
"score_path": "/workspace/dnathinker/runs/eval_reasoning_t3_v7r128_postRL_alpha1_s3_par_20260507_005010/score.json",
|
| 82 |
+
"tfg": 0.19508590161534112,
|
| 83 |
+
"n_cited": 15.08,
|
| 84 |
+
"n_grounded": 5.22,
|
| 85 |
+
"n_halluc": 9.54,
|
| 86 |
+
"reasoning_tags_rate": 0.8
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"seed": 42,
|
| 90 |
+
"score_path": "/workspace/dnathinker/runs/eval_reasoning_t3_v7r128_postRL_alpha1_20260506_004903/score.json",
|
| 91 |
+
"tfg": 0.23932521613754584,
|
| 92 |
+
"n_cited": 20.06,
|
| 93 |
+
"n_grounded": 7.8,
|
| 94 |
+
"n_halluc": 11.88,
|
| 95 |
+
"reasoning_tags_rate": 0.64
|
| 96 |
+
}
|
| 97 |
+
],
|
| 98 |
+
"tfg_stats": {
|
| 99 |
+
"n": 3,
|
| 100 |
+
"mean": 0.22470685776714827,
|
| 101 |
+
"std": 0.025653219924440397,
|
| 102 |
+
"min": 0.19508590161534112,
|
| 103 |
+
"max": 0.23970945554855788,
|
| 104 |
+
"values": [
|
| 105 |
+
0.23970945554855788,
|
| 106 |
+
0.19508590161534112,
|
| 107 |
+
0.23932521613754584
|
| 108 |
+
]
|
| 109 |
+
}
|
| 110 |
+
}
|
| 111 |
+
}
|