Refusal Rate Evaluation on XSTest Safe (test)
0Refusal RateSelf-CD
Evaluation Results
| Method | Links | |
|---|---|---|
| Self-CDModel=Mistral-7B2024.01 | 0 | |
| Self-CDModel=Vicuna-7B2024.01 | 2 | |
| Self-CDModel=InternLM-7B2024.01 | 2 | |
| Self-CDModel=Beaver2024.01 | 2.2 | |
| Self-CDModel=Vicuna-13B2024.01 | 3.2 | |
| NoSystemModel=Vicuna-7B2024.01 | 4 | |
| NoSystemModel=Vicuna-13B2024.01 | 4.4 | |
| NoSystemModel=Mistral-7B2024.01 | 4.4 | |
| CoT(zero)Model=Mistral-7B2024.01 | 5.6 | |
| NoSystemModel=Beaver2024.01 | 6.8 | |
| CoT(zero)Model=Vicuna-13B2024.01 | 7.6 | |
| CoT(zero)Model=Vicuna-7B2024.01 | 8 | |
| ICLModel=Beaver2024.01 | 8.4 | |
| Self-CDModel=LLaMA2-13B2024.01 | 9.6 | |
| Self-CDModel=LLaMA2-70B2024.01 | 9.6 | |
| CoT(One)Model=Vicuna-13B2024.01 | 9.6 | |
| ICLModel=Mistral-7B2024.01 | 9.6 | |
| PromptModel=InternLM-7B2024.01 | 9.6 | |
| Self-CDModel=LLaMA2-7B2024.01 | 10 | |
| PromptModel=Vicuna-7B2024.01 | 10 | |
| PromptModel=Mistral-7B2024.01 | 10 | |
| CoT(One)Model=Mistral-7B2024.01 | 11.2 | |
| CoT(One)Model=Vicuna-7B2024.01 | 11.6 | |
| ICLModel=Vicuna-7B2024.01 | 11.6 | |
| CoT(One)Model=InternLM-7B2024.01 | 12 | |
| CoT(zero)Model=Beaver2024.01 | 12.4 | |
| SystemModel=Vicuna-13B2024.01 | 14 | |
| ICLModel=Vicuna-13B2024.01 | 14 | |
| PromptModel=Beaver2024.01 | 15.8 | |
| PromptModel=Vicuna-13B2024.01 | 17.2 | |
| CoT(zero)Model=InternLM-7B2024.01 | 18 | |
| SystemModel=Beaver2024.01 | 18.4 | |
| SystemModel=Vicuna-7B2024.01 | 18.8 | |
| CoT(One)Model=Beaver2024.01 | 20.4 | |
| CoT(One)Model=LLaMA2-13B2024.01 | 30 | |
| PromptModel=LLaMA2-70B2024.01 | 31.2 | |
| CoT(One)Model=LLaMA2-70B2024.01 | 31.2 | |
| NoSystemModel=LLaMA2-13B2024.01 | 32 | |
| PromptModel=LLaMA2-13B2024.01 | 32.4 | |
| PromptModel=LLaMA2-7B2024.01 | 34.8 | |
| CoT(zero)Model=LLaMA2-13B2024.01 | 36 | |
| CoT(zero)Model=LLaMA2-7B2024.01 | 37.6 | |
| NoSystemModel=LLaMA2-7B2024.01 | 38 | |
| CoT(zero)Model=LLaMA2-70B2024.01 | 38 | |
| NoSystemModel=InternLM-7B2024.01 | 39.2 | |
| ICLModel=LLaMA2-7B2024.01 | 39.6 | |
| NoSystemModel=LLaMA2-70B2024.01 | 39.6 | |
| CoT(One)Model=LLaMA2-7B2024.01 | 41.6 | |
| ICLModel=LLaMA2-13B2024.01 | 42 | |
| SystemModel=LLaMA2-13B2024.01 | 42.4 | |
| ICLModel=InternLM-7B2024.01 | 42.5 | |
| SystemModel=Mistral-7B2024.01 | 44.8 | |
| SystemModel=InternLM-7B2024.01 | 44.8 | |
| ICLModel=LLaMA2-70B2024.01 | 50.8 | |
| SystemModel=LLaMA2-7B2024.01 | 54.4 | |
| SystemModel=LLaMA2-70B2024.01 | 59.6 |