Safety Evaluation on Safety Overall
32.9Reasoning Accuracy (Avg)No Defense
Evaluation results
| Method | Links | ||
|---|---|---|---|
| No DefenseModel=R1-Qwen-7B2025.08 | 32.9 | 46.4 | |
| No DefenseModel=R1-Llama-8B2025.08 | 26.1 | 41.8 | |
| ReasoningGuardModel=R1-Llama-8B2025.08 | 0.4 | 5.7 | |
| ReasoningGuardModel=R1-Qwen-7B2025.08 | 0.3 | 4.6 |