Safety Detection on Polyguard HR
0.9982AUROCT3
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| T3classifier=OCSVM2026.02 | 0.9982 | 0.0039 | |
| T3classifier=GMM2026.02 | 0.9965 | 0.0062 | |
| CIDER2026.02 | 0.8299 | 0.9059 | |
| LlamaGuard3parameters=1B, output_format=logits2026.02 | 0.8257 | 0.7417 | |
| MDJudge2026.02 | 0.7146 | 0.8807 | |
| GMM2026.02 | 0.703 | 0.9826 | |
| RMD2026.02 | 0.6864 | 0.9484 | |
| WildGuard2026.02 | 0.6833 | 1 | |
| LlamaGuard3parameters=1B2026.02 | 0.6368 | 1 | |
| DuoGuard2026.02 | 0.6363 | 0.9909 | |
| FDBD2026.02 | 0.6141 | 0.9951 | |
| LlamaGuard4parameters=12B2026.02 | 0.552 | 1 | |
| VIM2026.02 | 0.5153 | 0.999 | |
| REACT2026.02 | 0.5151 | 0.995 | |
| NNGUIDE2026.02 | 0.5142 | 0.9974 | |
| ADASCALE2026.02 | 0.467 | 0.9978 | |
| OPENMAX2026.02 | 0.4304 | 0.9967 | |
| Polyguard2026.02 | 0.4198 | 0.9484 |