List your company

Refusal Rate Evaluation on XSTest Safe (test)

0Refusal RateSelf-CD
-2.38413.70829.845.892Jan 31, 2024

Evaluation results

MethodLinks
Self-CDModel=Mistral-7B2024.010
Self-CDModel=Vicuna-7B2024.012
Self-CDModel=InternLM-7B2024.012
Self-CDModel=Beaver2024.012.2
Self-CDModel=Vicuna-13B2024.013.2
NoSystemModel=Vicuna-7B2024.014
NoSystemModel=Vicuna-13B2024.014.4
NoSystemModel=Mistral-7B2024.014.4
CoT(zero)Model=Mistral-7B2024.015.6
NoSystemModel=Beaver2024.016.8
CoT(zero)Model=Vicuna-13B2024.017.6
CoT(zero)Model=Vicuna-7B2024.018
ICLModel=Beaver2024.018.4
Self-CDModel=LLaMA2-13B2024.019.6
Self-CDModel=LLaMA2-70B2024.019.6
CoT(One)Model=Vicuna-13B2024.019.6
ICLModel=Mistral-7B2024.019.6
PromptModel=InternLM-7B2024.019.6
Self-CDModel=LLaMA2-7B2024.0110
PromptModel=Vicuna-7B2024.0110
PromptModel=Mistral-7B2024.0110
CoT(One)Model=Mistral-7B2024.0111.2
CoT(One)Model=Vicuna-7B2024.0111.6
ICLModel=Vicuna-7B2024.0111.6
CoT(One)Model=InternLM-7B2024.0112
CoT(zero)Model=Beaver2024.0112.4
SystemModel=Vicuna-13B2024.0114
ICLModel=Vicuna-13B2024.0114
PromptModel=Beaver2024.0115.8
PromptModel=Vicuna-13B2024.0117.2
CoT(zero)Model=InternLM-7B2024.0118
SystemModel=Beaver2024.0118.4
SystemModel=Vicuna-7B2024.0118.8
CoT(One)Model=Beaver2024.0120.4
CoT(One)Model=LLaMA2-13B2024.0130
PromptModel=LLaMA2-70B2024.0131.2
CoT(One)Model=LLaMA2-70B2024.0131.2
NoSystemModel=LLaMA2-13B2024.0132
PromptModel=LLaMA2-13B2024.0132.4
PromptModel=LLaMA2-7B2024.0134.8
CoT(zero)Model=LLaMA2-13B2024.0136
CoT(zero)Model=LLaMA2-7B2024.0137.6
NoSystemModel=LLaMA2-7B2024.0138
CoT(zero)Model=LLaMA2-70B2024.0138
NoSystemModel=InternLM-7B2024.0139.2
ICLModel=LLaMA2-7B2024.0139.6
NoSystemModel=LLaMA2-70B2024.0139.6
CoT(One)Model=LLaMA2-7B2024.0141.6
ICLModel=LLaMA2-13B2024.0142
SystemModel=LLaMA2-13B2024.0142.4
ICLModel=InternLM-7B2024.0142.5
SystemModel=Mistral-7B2024.0144.8
SystemModel=InternLM-7B2024.0144.8
ICLModel=LLaMA2-70B2024.0150.8
SystemModel=LLaMA2-7B2024.0154.4
SystemModel=LLaMA2-70B2024.0159.6