False Refusal Evaluation on ORB
90ScoreVCL
Evaluation Results
| Method | Links | |
|---|---|---|
| VCLBase Model=Llama-3.1-8B2026.03 | 90 | |
| Llama-3.1-8B-SFTBase Model=Llama-3.1-8B2026.03 | 80 | |
| DROBase Model=Llama-3.1-8B2026.03 | 74 | |
| System PromptBase Model=Llama-3.1-8B2026.03 | 69 | |
| Vector AblationBase Model=Llama-3.1-8B2026.03 | 53 | |
| Self-CDBase Model=Llama-3.1-8B2026.03 | 50 |