Safety Evaluation on UnsafeDiff (test)
94.2F1 ScoreMultiheaded
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Multiheadedopensourced=false2025.01 | 94.2 | — | |
| ImageGuard2025.01 | 68.9 | 0.808 | |
| GPT-4o2025.01 | 62.5 | — | |
| LlavaGuard2025.01 | 53 | — | |
| PerspectiveVisionopensourced=false2025.01 | 50 | — | |
| Claude3.5-sonnet2025.01 | 48.9 | — | |
| InternVL22025.01 | 47.7 | — | |
| Gemini1.5-pro2025.01 | 37.9 | — | |
| SD_filter2025.01 | 35.8 | — | |
| Idefics22025.01 | 32.5 | — | |
| LlavaNext2025.01 | 24 | — |