Refusal Evaluation on Do-Not-Answer
95.21Refusal RateLow-Rank Combination
Evaluation Results
| Method | Links | |
|---|---|---|
| Low-Rank CombinationBackbone=REFUSE-LLAMA, Method=Low-Rank Combination2026.03 | 95.21 | |
| Categorical SteeringBackbone=REFUSE-LLAMA, Method=Categorical Steering2026.03 | 92.55 | |
| REFUSE-LLAMABackbone=REFUSE-LLAMA2026.03 | 87.01 | |
| Low-Rank CombinationBackbone=DEEPSEEK R1 DISTILL LLAMA, Method=Low-Rank Combination2026.03 | 71.67 | |
| DEEPSEEK R1 DISTILL LLAMABackbone=DEEPSEEK R1 DISTILL LLAMA2026.03 | 69.86 | |
| LLAMA 3 8B INSTRUCTBackbone=LLAMA 3 8B INSTRUCT2026.03 | 54.85 | |
| Low-Rank CombinationBackbone=LLAMA 3 8B INSTRUCT, Method=Low-Rank Combination2026.03 | 50.48 |