Jailbreak Detection on JailbreakV
99.69AUROCMahal-OOD
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Mahal-OODModel=LLaVA, Uses target model hidden states=true2025.12 | 99.69 | 99.11 | |
| Mahal-OOD*Model=FLAVA, Uses target model hidden states=false2025.12 | 99.06 | 97.25 | |
| KNN-OODModel=LLaVA, Uses target model hidden states=true2025.12 | 98.33 | 95.88 | |
| KNN-OOD*Model=FLAVA, Uses target model hidden states=false2025.12 | 95 | 86.49 | |
| JailDAM*Model=CLIP, Uses target model hidden states=false2025.12 | 94.65 | 94.64 | |
| GradSafeModel=LLaVA2025.12 | 90.82 | 88.16 | |
| LLaVaGuardModel=Qwen2025.12 | 84.26 | 85.89 | |
| HiddenDetectModel=LLaVA2025.12 | 83.3 | 87.7 | |
| VLGuardModel=LLaVA2025.12 | 60.72 | 64.74 |