Safety Moderation on Safe RLHF AR
92F1 ScoreFanarGuard
Evaluation Results
| Method | Links | |
|---|---|---|
| FanarGuardSize=4B2026.03 | 92 | |
| PolyGuard-MinistralSize=8B2026.03 | 91 | |
| PolyGuard-QwenSize=7B2026.03 | 90 | |
| Llama-Guard-3Size=8B2026.03 | 84 | |
| PolyGuard-SmolSize=0.5B2026.03 | 82 | |
| ShieldGemma-2bSize=2B2026.03 | 75 | |
| WildguardSize=7B2026.03 | 65 | |
| MD-JudgeSize=7B2026.03 | 32 |