| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Overrefusal Evaluation | OrBench-H | RR99.85 | 21 | |
| Safety Boundary Over-Refusal | ORBench | ORBench Over-Refusal Rate7.2 | 20 | |
| Safety and Informativeness Evaluation | OrBench Hard | Deception Rate (Safe)75 | 4 | |
| Safety Evaluation | OrBench Toxic | Safety Score81.7 | 4 |