| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| Safety Evaluation Suite (AEG, AEG2, HB-A, HB-V, OAI, SAL, SST, Toxic, WGT, WJB) | WildGuard | AEG Score89.8 | 18 | 26d ago | |
| JailbreakBench Conv | ReGA | MCA Accuracy100 | 7 | 3mo ago | |
| RepE-Harmful | ReGA | MCA Accuracy100 | 7 | 3mo ago | |
| JailbreakBench | ReGA | Accuracy (MCA)100 | 7 | 3mo ago | |
| HarmBench | ReGA | MCA Accuracy100 | 7 | 3mo ago |
