List your company

Safety Evaluation on XSTest (F1)

97F1 ScoreSafeKey
8.631.5554.577.45Aug 6, 2025

Evaluation results

MethodLinks
SafeKeyModel=R1-Llama-8B, Output Component=Final Answer2025.0897
ReasoningGuardModel=R1-Llama-8B, Output Component=Reasoning2025.0897
SafeKeyModel=R1-Llama-8B, Output Component=Reasoning2025.0895
Self-ReminderModel=R1-Qwen-7B, Output Component=Reasoning2025.0894
Self-ReminderModel=R1-Llama-8B, Output Component=Reasoning2025.0893
SafeKeyModel=R1-Qwen-7B, Output Component=Reasoning2025.0892
Self-ReminderModel=R1-Qwen-7B, Output Component=Final Answer2025.0889
SafeDecodingModel=R1-Llama-8B, Output Component=Reasoning2025.0888
RealSafe-R1Model=R1-Llama-8B, Output Component=Reasoning2025.0888
SafeDecodingModel=R1-Qwen-7B, Output Component=Reasoning2025.0887
SafeKeyModel=R1-Qwen-7B, Output Component=Final Answer2025.0887
RealSafe-R1Model=R1-Qwen-7B, Output Component=Reasoning2025.0886
No DefenseModel=R1-Llama-8B, Output Component=Reasoning2025.0886
SAFEPATH-ZSModel=R1-Qwen-7B, Output Component=Reasoning2025.0885
Self-ReminderModel=R1-Llama-8B, Output Component=Final Answer2025.0885
ReasoningGuardModel=R1-Llama-8B, Output Component=Final Answer2025.0885
No DefenseModel=R1-Qwen-7B, Output Component=Reasoning2025.0884
ParaphraseModel=R1-Qwen-7B, Output Component=Reasoning2025.0884
ParaphraseModel=R1-Llama-8B, Output Component=Reasoning2025.0883
No DefenseModel=R1-Qwen-7B, Output Component=Final Answer2025.0881
RealSafe-R1Model=R1-Llama-8B, Output Component=Final Answer2025.0881
SafeDecodingModel=R1-Qwen-7B, Output Component=Final Answer2025.0880
SAFEPATH-ZSModel=R1-Qwen-7B, Output Component=Final Answer2025.0880
ReasoningGuardModel=R1-Qwen-7B, Output Component=Reasoning2025.0880
SafeDecodingModel=R1-Llama-8B, Output Component=Final Answer2025.0880
SAFEPATH-ZSModel=R1-Llama-8B, Output Component=Final Answer2025.0880
SmoothLLMModel=R1-Qwen-7B, Output Component=Reasoning2025.0879
SmoothLLMModel=R1-Llama-8B, Output Component=Reasoning2025.0879
No DefenseModel=R1-Llama-8B, Output Component=Final Answer2025.0878
ParaphraseModel=R1-Llama-8B, Output Component=Final Answer2025.0878
ParaphraseModel=R1-Qwen-7B, Output Component=Final Answer2025.0877
SmoothLLMModel=R1-Qwen-7B, Output Component=Final Answer2025.0877
SAFEPATH-ZSModel=R1-Llama-8B, Output Component=Reasoning2025.0877
SmoothLLMModel=R1-Llama-8B, Output Component=Final Answer2025.0876
ThinkingIModel=R1-Llama-8B, Output Component=Final Answer2025.0874
SAFEPATH-FTModel=R1-Qwen-7B, Output Component=Final Answer2025.0873
ReasoningGuardModel=R1-Qwen-7B, Output Component=Final Answer2025.0864
RealSafe-R1Model=R1-Qwen-7B, Output Component=Final Answer2025.0863
SAFEPATH-FTModel=R1-Qwen-7B, Output Component=Reasoning2025.0862
ThinkingIModel=R1-Qwen-7B, Output Component=Final Answer2025.0858
SAFEPATH-FTModel=R1-Llama-8B, Output Component=Final Answer2025.0856
SAFEPATH-FTModel=R1-Llama-8B, Output Component=Reasoning2025.0850
ThinkingIModel=R1-Qwen-7B, Output Component=Reasoning2025.0821
ThinkingIModel=R1-Llama-8B, Output Component=Reasoning2025.0812