Safety Evaluation on HEx-PHI-FigStep
1Safety Score (1-ASR)PSI
Evaluation Results
| Method | Links | |
|---|---|---|
| PSIBackbone Model=HuatuoGPT-Vision-7B2025.12 | 1 | |
| RESTABackbone Model=Lingshu-7B2025.12 | 0.9982 | |
| RESTABackbone Model=HuatuoGPT-Vision-7B2025.12 | 0.9927 | |
| PSIBackbone Model=Lingshu-7B2025.12 | 0.94 | |
| OriginalBackbone Model=HuatuoGPT-Vision-7B2025.12 | 0.28 | |
| OriginalBackbone Model=Lingshu-7B2025.12 | 0.25 | |
| ModelMergeBackbone Model=Lingshu-7B2025.12 | 0.2164 | |
| ModelMergeBackbone Model=HuatuoGPT-Vision-7B2025.12 | 0.1673 |