| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| Nectar | AAD | Reward (R)3.7 | 30 | 1mo ago | |
| Pairwise | CFA | Average Score92.12 | 18 | 4mo ago | |
| Indomain | Vicuna-7B | Win Rate0 | 11 | 4mo ago | |
| PRISM (test) | Hard Panel | BT Score (Mean)0.331 | 10 | 4mo ago | |
| PKU-SafeRLHF | Pref-CTRL | Win Rate (Llama Judge)83 | 4 | 2mo ago |