List your company

Jailbreak Detection on AdvBench Llama2-7B

97.9AUROCGradSafe
14.80436.37757.9579.523Jun 5, 2026

Evaluation results

MethodLinks
GradSafeAttack=AutoDAN2026.0697.9
MTKAttack=SAA2026.0696.1
MTKAttack=DrAttack2026.0695.9
MTKAttack=TAP2026.0695.6
MTKAttack=PAIR2026.0695.3
SelfDefendAttack=SAA2026.0694.6
SelfDefendAttack=AutoDAN2026.0694.5
MTKAttack=PAP2026.0694.4
GradSafeAttack=TAP2026.0694.1
MTKAttack=Avg.2026.0694
SelfDefendAttack=GCG2026.0693.9
GradSafeAttack=PAIR2026.0693.6
HSFAttack=DrAttack2026.0692.9
MTKAttack=JailJudge2026.0692.9
MTKAttack=GCG2026.0692.9
MTKAttack=AutoDAN2026.0692.7
HiddenDetectAttack=AutoDAN2026.0692.7
MTKAttack=IJP2026.0692.4
HiddenDetectAttack=JailJudge2026.0692.4
MTKAttack=Zulu2026.0692.2
HiddenDetectAttack=Zulu2026.0692.1
GradSafeAttack=GCG2026.0690.4
SelfDefendAttack=PAIR2026.0689.3
SelfDefendAttack=IJP2026.0689.1
SaPAttack=Zulu2026.0689
GradSafeAttack=SAA2026.0688.6
HSFAttack=IJP2026.0687.6
SelfDefendAttack=Zulu2026.0687.6
GradSafeAttack=IJP2026.0686.3
SelfDefendAttack=TAP2026.0686.1
HiddenDetectAttack=TAP2026.0685.8
GradSafeAttack=JailJudge2026.0685.7
HiddenDetectAttack=PAIR2026.0685.6
HSFAttack=TAP2026.0685.3
SelfDefendAttack=Avg.2026.0684.8
SaPAttack=PAP2026.0684.4
SmoothLLMAttack=JailJudge2026.0683.5
HSFAttack=SAA2026.0682.5
GradSafeAttack=Avg.2026.0681.3
SelfDefendAttack=PAP2026.0680.1
HSFAttack=PAIR2026.0679
HiddenDetectAttack=GCG2026.0678.6
SaPAttack=AutoDAN2026.0678
GradSafeAttack=DrAttack2026.0677.6
HiddenDetectAttack=Avg.2026.0677.6
GradCuffAttack=TAP2026.0677.2
HiddenDetectAttack=DrAttack2026.0676.3
GradCuffAttack=AutoDAN2026.0675.8
SmoothLLMAttack=Zulu2026.0674.5
SmoothLLMAttack=AutoDAN2026.0673.3
GradSafeAttack=Zulu2026.0673.3
HSFAttack=Avg.2026.0672
GradCuffAttack=PAIR2026.0671.9
SmoothLLMAttack=SAA2026.0671.8
SelfDefendAttack=DrAttack2026.0671.3
GradCuffAttack=IJP2026.0671.2
HSFAttack=GCG2026.0670.2
HSFAttack=PAP2026.0669.5
SaPAttack=IJP2026.0668.5
HiddenDetectAttack=IJP2026.0667.3
GradCuffAttack=JailJudge2026.0667.3
SaPAttack=GCG2026.0665.3
GradCuffAttack=Avg.2026.0664.7
GradCuffAttack=SAA2026.0663.7
SmoothLLMAttack=PAP2026.0663.4
SaPAttack=Avg.2026.0663.3
SelfDefendAttack=JailJudge2026.0661.9
HiddenDetectAttack=SAA2026.0661.5
GradCuffAttack=DrAttack2026.0657.6
GradCuffAttack=GCG2026.0657
HSFAttack=JailJudge2026.0656.1
SmoothLLMAttack=Avg.2026.0655.8
SmoothLLMAttack=GCG2026.0655.1
GradCuffAttack=Zulu2026.0654.5
HSFAttack=AutoDAN2026.0651.8
SmoothLLMAttack=TAP2026.0651.6
GradCuffAttack=PAP2026.0651.2
SaPAttack=PAIR2026.0650.9
SaPAttack=TAP2026.0650.2
SaPAttack=DrAttack2026.0650
SaPAttack=JailJudge2026.0649.7
SaPAttack=SAA2026.0646.6
HSFAttack=Zulu2026.0644.8
SmoothLLMAttack=PAIR2026.0644.4
HiddenDetectAttack=PAP2026.0643.2
GradSafeAttack=PAP2026.0625.2
SmoothLLMAttack=IJP2026.0622.4
SmoothLLMAttack=DrAttack2026.0618