List your company
ResearchFormerly WizWand

Jailbreak Attack on Safety Prompt Evaluation Set

0.8ASRVanilla
-2.36418.99340.3561.707Sep 15, 2025
Updated 7mo ago

Evaluation results

MethodLinks
VanillaTarget Model=Gemini-2.0-Flash2025.090.8
VanillaTarget Model=Gemini-2.0-Flash-Lite2025.091
VanillaTarget Model=Gemini-1.5-Pro2025.091.4
VanillaTarget Model=QwQ-32B2025.093.6
JailbreakTarget Model=Gemini-1.5-Pro2025.095.3
JailbreakTarget Model=QwQ-32B2025.096.3
VanillaTarget Model=Gemma-3-1b-it2025.0910.6
JailbreakTarget Model=Gemini-2.0-Flash-Lite2025.0915.6
JailbreakTarget Model=Gemini-2.0-Flash2025.0915.7
JailbreakTarget Model=Gemma-3-1b-it2025.0924.6
NeuroStrikeTarget Model=Gemini-2.0-Flash-Lite2025.0949.2
NeuroStrikeTarget Model=Gemini-2.0-Flash2025.0954.7
NeuroStrikeTarget Model=Gemini-1.5-Pro2025.0955.7
NeuroStrikeTarget Model=QwQ-32B2025.0978.9
NeuroStrikeTarget Model=Gemma-3-1b-it2025.0979.9