| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| HarmfulQA | Aligner | Helpfulness Score69.4 | 33 | 4d ago | |
| Beavertails | Aligner | Helpful Score58.4 | 33 | 4d ago | |
| SafeEdit | DINM | Success Rate99.26 | 16 | 3d ago | |
| HH-RLHF harmless (test) | APL | Win Rate83.33 | 12 | 3d ago | |
| VLSafe (test) | LLaVA-HF | Relevance100 | 7 | 3d ago | |
| HarmfulQ (test) | DeAL | Harmlessness Fraction100 | 7 | 4d ago | |
| Harmlessness | MOPO | Disc. Score0.5409 | 5 | 4d ago |