| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| Braceval | Accuracy70.8 | 15 | 4mo ago | ||
| CRSArena Eval (All) | FACE | Pearson Correlation (r)0.918 | 9 | 1mo ago | |
| CRSArena-Eval Dial-level | FACE | Pearson r0.913 | 9 | 1mo ago | |
| CRSArena-Eval Turn-level | FACE | Pearson Correlation (r)0.93 | 9 | 1mo ago | |
| Grocery Domain | GOOD | Human Score (%)75.78 | 3 | 2mo ago | |
| Robot Domain | GOOD | Human Score88.57 | 3 | 2mo ago | |
| Online Chinese conversation data 500 multi-turn proprietary dialogues (test) | C-SFT-Empathy | Win Rate vs GPT479.4 | 2 | 2mo ago |
