| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| WildGuardMix (test) | D2-MLP | Accuracy89.9 | 40 | 8d ago | |
| UAV dataset | Precision92.5 | 15 | 2mo ago | ||
| AutoDAN | ActWM | ASR @ 1% FPR67.86 | 5 | 2mo ago | |
| Multilingual | ActWM | ASR @ 1% FPR34.15 | 5 | 2mo ago | |
| DeepInception | ActWM | ASR @ 1% FPR67.02 | 5 | 2mo ago | |
| Jailbroken | ASR @ 1% FPR32.96 | 5 | 2mo ago | ||
| BeaverTails | QwenGuard | ASR @ 1% FPR19.6 | 5 | 2mo ago | |
| COCO-O Avg | KGFP | Safety AUROC80.3 | 4 | 2mo ago | |
| COCO (val) | KGFP | Safety AUROC92.9 | 4 | 2mo ago |