OOD safety category inference (Stage 2) on ProGuard Text-Image
26.86Mean RewardProGuard-7B
Evaluation Results
| Method | Links | |
|---|---|---|
| ProGuard-7B2025.12 | 26.86 | |
| ProGuard-3B2025.12 | 20.89 | |
| Gemini2.5-Flash2025.12 | 4.43 | |
| GPT4o-mini2025.12 | 0 |
Loading the SOTA2 catalog…
| Method | Links | |
|---|---|---|
| ProGuard-7B2025.12 | 26.86 | |
| ProGuard-3B2025.12 | 20.89 | |
| Gemini2.5-Flash2025.12 | 4.43 | |
| GPT4o-mini2025.12 | 0 |