Safety and Helpfulness Evaluation on MM-Safety
88.71Safety ScoreSAFER-STEER
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SAFER-STEERBase Model=Qwen2.5VL-7B, Method Category=Dynamic Methods2026.03 | 88.71 | 73.42 | |
| SFTBase Model=Qwen2.5VL-7B, Method Category=Dynamic Methods2026.03 | 83.45 | 67.09 | |
| SAFER-STEERBase Model=Qwen2.5VL-3B, Method Category=Dynamic Methods2026.03 | 81.08 | 63.56 | |
| SFTBase Model=Qwen2.5VL-3B, Method Category=Dynamic Methods2026.03 | 80.7 | 61.51 | |
| TISBase Model=Qwen2.5VL-7B, Method Category=Static Methods2026.03 | 79.11 | 70 | |
| SaFeR-VLMBase Model=Qwen2.5VL-3B, Method Category=Static Methods2026.03 | 74.52 | 61.79 | |
| TISBase Model=Qwen2.5VL-3B, Method Category=Static Methods2026.03 | 72.79 | 61.52 | |
| SPA-VLBase Model=Qwen2.5VL-3B, Method Category=Static Methods2026.03 | 72.45 | 57.85 | |
| SaFeR-VLMBase Model=Qwen2.5VL-7B, Method Category=Static Methods2026.03 | 72.08 | 62.89 | |
| SPA-VLBase Model=Qwen2.5VL-7B, Method Category=Static Methods2026.03 | 68.94 | 51.64 | |
| MMDPOBase Model=Qwen2.5VL-3B, Method Category=Static Methods2026.03 | 63.55 | 54.16 | |
| MMDPOBase Model=Qwen2.5VL-7B, Method Category=Static Methods2026.03 | 57.9 | 46.65 | |
| GRLHF-VBase Model=Qwen2.5VL-7B, Method Category=Static Methods2026.03 | 46.68 | 33.82 | |
| Qwen2.5VL-3BBase Model=Qwen2.5VL-3B, Method Category=Base2026.03 | 36.1 | 25.2 | |
| SafeRLHF-VBase Model=Qwen2.5VL-7B, Method Category=Static Methods2026.03 | 35.18 | 21.96 | |
| GRLHF-VBase Model=Qwen2.5VL-3B, Method Category=Static Methods2026.03 | 34.77 | 27.39 | |
| SafeRLHF-VBase Model=Qwen2.5VL-3B, Method Category=Static Methods2026.03 | 22.22 | 9.86 | |
| Qwen2.5VL-7BBase Model=Qwen2.5VL-7B, Method Category=Base2026.03 | 20.94 | 37.72 |