OOD safety category inference (Stage 2) on ProGuard Image
25.95Mean RewardProGuard-3B
Evaluation Results
| Method | Links | |
|---|---|---|
| ProGuard-3B2025.12 | 25.95 | |
| ProGuard-7B2025.12 | 18.76 | |
| Gemini2.5-Flash2025.12 | 4.57 | |
| GPT4o-mini2025.12 | 0.02 |
Loading the SOTA2 catalog…
| Method | Links | |
|---|---|---|
| ProGuard-3B2025.12 | 25.95 | |
| ProGuard-7B2025.12 | 18.76 | |
| Gemini2.5-Flash2025.12 | 4.57 | |
| GPT4o-mini2025.12 | 0.02 |