List your company

Jailbreaking on NeuroStrike Evaluation Prompts

43.2Attack Success Rate (ASR)NeuroStrike
41.73651.61861.571.382Sep 15, 2025

Evaluation results

MethodLinks
NeuroStrikeTarget Model=Gemini-2.0-Flash-Lite, Defense Method=Perplexity Filter2025.0943.2
NeuroStrikeTarget Model=QwQ-32B, Defense Method=LSE2025.0943.4
NeuroStrikeTarget Model=Gemini-2.0-Flash-Lite, Defense Method=SmoothLLM2025.0947.3
NeuroStrikeTarget Model=Gemini-2.0-Flash, Defense Method=Perplexity Filter2025.0948.7
NeuroStrikeTarget Model=Gemini-1.5-Pro, Defense Method=Perplexity Filter2025.0949.7
NeuroStrikeTarget Model=Gemini-2.0-Flash, Defense Method=SmoothLLM2025.0952.8
NeuroStrikeTarget Model=Gemini-1.5-Pro, Defense Method=SmoothLLM2025.0953.8
NeuroStrikeTarget Model=Gemma-3-1b-it, Defense Method=LSE2025.0960
NeuroStrikeTarget Model=QwQ-32B, Defense Method=SmoothLLM2025.0976.4
NeuroStrikeTarget Model=Gemma-3-1b-it, Defense Method=SmoothLLM2025.0978
NeuroStrikeTarget Model=QwQ-32B, Defense Method=Perplexity Filter2025.0978.8
NeuroStrikeTarget Model=Gemma-3-1b-it, Defense Method=Perplexity Filter2025.0979.8