OOD safety category inference (Stage 2) on LlavaGuard
13.28Reward MeanGemini2.5-Flash
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini2.5-Flash2025.12 | 13.28 | |
| ProGuard-7B2025.12 | 5.98 | |
| ProGuard-3B2025.12 | 5.31 | |
| GPT4o-mini2025.12 | 0.71 |
Loading the SOTA2 catalog…
| Method | Links | |
|---|---|---|
| Gemini2.5-Flash2025.12 | 13.28 | |
| ProGuard-7B2025.12 | 5.98 | |
| ProGuard-3B2025.12 | 5.31 | |
| GPT4o-mini2025.12 | 0.71 |