Prompt Safety Detection on WildGuardMix (train)
0.8971AUROCT3+GMM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| T3+GMMDetection Head=GMM2026.02 | 0.8971 | 0.2422 | |
| T3+OCSVMDetection Head=OCSVM2026.02 | 0.8853 | 0.2802 | |
| POLYGUARD2026.02 | 0.8721 | 0.4686 | |
| MDJUDGE2026.02 | 0.8024 | 0.762 | |
| DUOGUARD2026.02 | 0.7661 | 0.8975 | |
| CIDER2026.02 | 0.7414 | 0.6586 | |
| LLAMAGUARD3-1BParameters=1B2026.02 | 0.7032 | 1 | |
| GMM2026.02 | 0.6809 | 0.873 | |
| OPENMAX2026.02 | 0.6077 | 0.9938 | |
| RMD2026.02 | 0.6018 | 0.9587 | |
| VIM2026.02 | 0.5241 | 0.9829 | |
| FDBD2026.02 | 0.5112 | 0.9707 | |
| NNGUIDE2026.02 | 0.4796 | 0.9849 | |
| ADASCALE2026.02 | 0.3259 | 0.9907 | |
| REACT2026.02 | 0.2662 | 0.9961 |