List your company

Jailbreak Attack on Grok 4

82NR ScoreAuto-DAN
-1.220.44263.6Mar 14, 2026

Evaluation results

MethodLinks
Auto-DANSurrogate Model=Qwen2.5-7B2026.038282
Auto-DANSurrogate Model=Mistral-7B2026.038281
Auto-DANSurrogate Model=Gemma-3-4B2026.037876.3
SWhisperSurrogate Model=LLaMA-3.1-8B2026.037873.3
SWhisperSurrogate Model=Gemma-3-4B2026.037869.5
SWhisperSurrogate Model=Qwen2.5-7B2026.037874
SWhisperSurrogate Model=Mistral-7B2026.037066
Auto-DANSurrogate Model=LLaMA-3.1-8B2026.036665.8
PAIRSurrogate Model=Qwen2.5-7B2026.036461.8
PAIRSurrogate Model=Gemma-3-4B2026.035047.5
PAIRSurrogate Model=LLaMA-3.1-8B2026.034643.8
PAIRSurrogate Model=Mistral-7B2026.034643.3
GCGSurrogate Model=LLaMA-3.1-8B2026.032220.8
GCGSurrogate Model=Gemma-3-4B2026.03109.8
IGCGSurrogate Model=LLaMA-3.1-8B2026.03109.3
IGCGSurrogate Model=Gemma-3-4B2026.0387.8
GCGSurrogate Model=Qwen2.5-7B2026.0366
GCGSurrogate Model=Mistral-7B2026.0366
IGCGSurrogate Model=Qwen2.5-7B2026.0366
IGCGSurrogate Model=Mistral-7B2026.0322