Safety on Cultural Kaleidoscope
76.6F1 ScoreDuoGuard
Evaluation Results
| Method | Links | |
|---|---|---|
| DuoGuardScale=Small (<=0.5B)2025.12 | 76.6 | |
| PG-QwenScale=Large (>=2.5B)2025.12 | 75.71 | |
| CREST-BASEScale=Small (<=0.5B)2025.12 | 69.42 | |
| WalledguardScale=Small (<=0.5B)2025.12 | 67.43 | |
| CREST-LARGEScale=Small (<=0.5B)2025.12 | 56.79 | |
| PQ-Qwen-SmolScale=Small (<=0.5B)2025.12 | 55.3 | |
| LlamaGuardScale=Large (>=2.5B)2025.12 | 26.87 |