Upload _reasoning_rl_multiseed/eval_reasoning_t3_v7r128_postRL_alpha1_s3_par_20260507_005010/score.md with huggingface_hub
Browse files
_reasoning_rl_multiseed/eval_reasoning_t3_v7r128_postRL_alpha1_s3_par_20260507_005010/score.md
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Reasoning quality metrics · n=50
|
| 2 |
+
|
| 3 |
+
## Format / sanity
|
| 4 |
+
|
| 5 |
+
| metric | value |
|
| 6 |
+
|---|---:|
|
| 7 |
+
| reasoning_tags_rate | 0.8000 |
|
| 8 |
+
| dna_tags_rate | 0.0000 |
|
| 9 |
+
| cell_mention_rate | 0.9200 |
|
| 10 |
+
| mean_len_chars | 1052.6200 |
|
| 11 |
+
| p10_len_chars | 735.4000 |
|
| 12 |
+
| p90_len_chars | 1656.0000 |
|
| 13 |
+
|
| 14 |
+
## TF-grounding (the headline reasoning metric)
|
| 15 |
+
|
| 16 |
+
| metric | value |
|
| 17 |
+
|---|---:|
|
| 18 |
+
| mean_tf_grounded_rate | 0.1951 |
|
| 19 |
+
| mean_n_cited | 15.0800 |
|
| 20 |
+
| mean_n_grounded | 5.2200 |
|
| 21 |
+
| mean_n_canonical_only | 0.3200 |
|
| 22 |
+
| mean_n_hallucinated | 9.5400 |
|
| 23 |
+
| mean_motif_consensus | 0.3223 |
|
| 24 |
+
| mean_prog_coverage | 0.1296 |
|
| 25 |
+
|
| 26 |
+
## Per-cell breakdown
|
| 27 |
+
|
| 28 |
+
| cell | n | tf_grounded_rate ↑ | n_cited | len_chars | cell_mention |
|
| 29 |
+
|---|---:|---:|---:|---:|---:|
|
| 30 |
+
| Ast | 5 | 0.179 | 9.8 | 1015 | 0.800 |
|
| 31 |
+
| End | 3 | 0.186 | 9.7 | 816 | 1.000 |
|
| 32 |
+
| Ex | 15 | 0.233 | 19.4 | 1017 | 0.867 |
|
| 33 |
+
| In | 4 | 0.086 | 12.5 | 1069 | 1.000 |
|
| 34 |
+
| Mic | 13 | 0.254 | 16.4 | 1167 | 0.923 |
|
| 35 |
+
| OPC | 2 | 0.216 | 11.5 | 937 | 1.000 |
|
| 36 |
+
| Oli | 8 | 0.091 | 12.4 | 1067 | 1.000 |
|