List your company

Harmfulness Evaluation on GCG

1.16Harmfulness ScoreSAFEPATH-FT
1.041.852.663.47Aug 6, 2025

Evaluation results

MethodLinks
SAFEPATH-FTModel=R1-Llama-8B2025.081.16
SAFEPATH-ZSModel=R1-Llama-8B2025.081.34
ReasoningGuardModel=R1-Llama-8B2025.081.34
SafeKeyModel=R1-Llama-8B2025.081.4
SAFEPATH-ZSModel=R1-Qwen-7B2025.081.42
RealSafe-R1Model=R1-Qwen-7B2025.081.44
SafeKeyModel=R1-Qwen-7B2025.081.44
ReasoningGuardModel=R1-Qwen-7B2025.081.46
SAFEPATH-FTModel=R1-Qwen-7B2025.081.54
RealSafe-R1Model=R1-Llama-8B2025.081.6
ThinkingIModel=R1-Qwen-7B2025.081.86
ThinkingIModel=R1-Llama-8B2025.081.96
Self-ReminderModel=R1-Llama-8B2025.081.98
Self-ReminderModel=R1-Qwen-7B2025.082.36
ParaphraseModel=R1-Llama-8B2025.082.56
ParaphraseModel=R1-Qwen-7B2025.082.86
SmoothLLMModel=R1-Llama-8B2025.083.08
SmoothLLMModel=R1-Qwen-7B2025.083.16
No DefenseModel=R1-Llama-8B2025.083.32
SafeDecodingModel=R1-Llama-8B2025.083.48
SafeDecodingModel=R1-Qwen-7B2025.083.82
No DefenseModel=R1-Qwen-7B2025.084.16