List your company
ResearchFormerly WizWand

Adversarial Jailbreak Attack on Vicuna 7B

98.46Attack Success Rate (ASR)SRA
1.656826.788451.9277.0516May 9, 2026
Updated 5mo ago

Evaluation results

MethodLinks
SRASource Model=Mistral 7B, Optimization Time=60s, Attack Setting=Black-box2026.0598.46——
SRASource Model=Mistral 7B, Optimization Time=15s, Attack Setting=Black-box2026.0598.27——
SRASource Model=Mistral 7B, Optimization Time=30s, Attack Setting=Black-box2026.0597.69——
SRABudget=60s2026.0597.518.6796.73
SRASource Model=Vicuna 7B, Optimization Time=60s, Attack Setting=White-box2026.0597.5——
SRABudget=30s2026.0596.9221.797.31
SRASource Model=Vicuna 7B, Optimization Time=30s, Attack Setting=White-box2026.0596.92——
SRABudget=15s2026.0595.7724.2195.96
SRASource Model=Vicuna 7B, Optimization Time=15s, Attack Setting=White-box2026.0595.77——
SRASource Model=Vicuna 13B, Optimization Time=30s, Attack Setting=Black-box2026.0594.42——
SRASource Model=Vicuna 13B, Optimization Time=60s, Attack Setting=Black-box2026.0594.42——
BEASTBudget=60s, Venue=ICML 20242026.0593.65156.9544.04
BEASTSource Model=Vicuna 7B, Optimization Time=60s, Attack Setting=White-box2026.0593.65——
SRASource Model=Vicuna 13B, Optimization Time=15s, Attack Setting=Black-box2026.0593.27——
SRASource Model=Guanaco 7B, Optimization Time=15s, Attack Setting=Black-box2026.0592.31——
SRASource Model=Guanaco 7B, Optimization Time=60s, Attack Setting=Black-box2026.0591.73——
SRASource Model=Guanaco 7B, Optimization Time=30s, Attack Setting=Black-box2026.0590.96——
BEASTBudget=30s, Venue=ICML 20242026.0590.19119.1563.85
BEASTSource Model=Vicuna 7B, Optimization Time=30s, Attack Setting=White-box2026.0590.19——
AutoDANSource Model=Mistral 7B, Optimization Time=60s, Attack Setting=Black-box2026.0582.12——
AutoDANSource Model=Mistral 7B, Optimization Time=15s, Attack Setting=Black-box2026.0581.73——
AutoDANSource Model=Mistral 7B, Optimization Time=30s, Attack Setting=Black-box2026.0581.35——
AutoDANSource Model=Guanaco 7B, Optimization Time=30s, Attack Setting=Black-box2026.0579.42——
AutoDANBudget=60s, Venue=ICLR 20242026.0579.0462.0777.12
AutoDANSource Model=Vicuna 7B, Optimization Time=60s, Attack Setting=White-box2026.0579.04——
AutoDANSource Model=Vicuna 13B, Optimization Time=30s, Attack Setting=Black-box2026.0578.65——
AutoDANBudget=30s, Venue=ICLR 20242026.0578.2761.2577.5
AutoDANSource Model=Vicuna 7B, Optimization Time=30s, Attack Setting=White-box2026.0578.27——
AutoDANSource Model=Vicuna 13B, Optimization Time=15s, Attack Setting=Black-box2026.0578.27——
AutoDANSource Model=Guanaco 7B, Optimization Time=15s, Attack Setting=Black-box2026.0577.88——
AutoDANSource Model=Guanaco 7B, Optimization Time=60s, Attack Setting=Black-box2026.0577.5——
BEASTBudget=15s, Venue=ICML 20242026.0577.1282.4767.31
AutoDANSource Model=Vicuna 7B, Optimization Time=15s, Attack Setting=White-box2026.0577.12——
AutoDANBudget=15s, Venue=ICLR 20242026.0575.1960.5578.27
BEASTSource Model=Vicuna 7B, Optimization Time=15s, Attack Setting=White-box2026.0575.19——
GCGBudget=60s, Venue=arXiv 20232026.0573.656,572.960
GCGSource Model=Vicuna 7B, Optimization Time=60s, Attack Setting=White-box2026.0573.65——
AutoDANSource Model=Vicuna 13B, Optimization Time=60s, Attack Setting=Black-box2026.0571.73——
GCGBudget=30s, Venue=arXiv 20232026.0561.153,741.860
GCGSource Model=Vicuna 7B, Optimization Time=30s, Attack Setting=White-box2026.0561.15——
BEASTSource Model=Vicuna 13B, Optimization Time=60s, Attack Setting=Black-box2026.0549.81——
GCGBudget=15s, Venue=arXiv 20232026.0543.85753.390.96
GCGSource Model=Vicuna 7B, Optimization Time=15s, Attack Setting=White-box2026.0543.85——
BEASTSource Model=Vicuna 13B, Optimization Time=30s, Attack Setting=Black-box2026.0542.88——
BEASTSource Model=Mistral 7B, Optimization Time=60s, Attack Setting=Black-box2026.0529.42——
BEASTSource Model=Guanaco 7B, Optimization Time=60s, Attack Setting=Black-box2026.0528.08——
BEASTSource Model=Vicuna 13B, Optimization Time=15s, Attack Setting=Black-box2026.0527.12——
BEASTSource Model=Mistral 7B, Optimization Time=30s, Attack Setting=Black-box2026.0524.42——
GCGSource Model=Guanaco 7B, Optimization Time=60s, Attack Setting=Black-box2026.0524.04——
GCGSource Model=Guanaco 7B, Optimization Time=30s, Attack Setting=Black-box2026.0520.96——
BEASTSource Model=Guanaco 7B, Optimization Time=30s, Attack Setting=Black-box2026.0519.81——
BEASTSource Model=Mistral 7B, Optimization Time=15s, Attack Setting=Black-box2026.0519.62——
GCGSource Model=Mistral 7B, Optimization Time=30s, Attack Setting=Black-box2026.0515.96——
GCGSource Model=Mistral 7B, Optimization Time=60s, Attack Setting=Black-box2026.0515——
GCGSource Model=Guanaco 7B, Optimization Time=15s, Attack Setting=Black-box2026.0514.62——
BEASTSource Model=Guanaco 7B, Optimization Time=15s, Attack Setting=Black-box2026.0514.62——
GCGSource Model=Mistral 7B, Optimization Time=15s, Attack Setting=Black-box2026.0510——
CleanBudget=-2026.055.3827.295.38