List your company

Jailbreak Defense on Original Dataset

5.82ASREDDF
3.19420.919538.64556.3705Feb 26, 2025

Evaluation results

MethodLinks
EDDFTarget Model=qwen plus2025.025.82
Defense PromptTarget Model=qwen plus2025.029.93
Intention AnalysisTarget Model=qwen plus2025.0212.58
Self-ReminderTarget Model=qwen plus2025.0216.37
EmbeddingTarget Model=qwen plus2025.0236.4
RegexTarget Model=qwen plus2025.0246.03
Llama3-GuardTarget Model=qwen plus2025.0255
Guard Few-shotTarget Model=qwen plus2025.0271.47