List your company

Jailbreak Detection on AdvBench

97.5Refusal RateLWV
-3.296822.871649.0475.2084Jun 3, 2026

Evaluation results

MethodLinks
LWVEvaluation Protocol=Zero-shot2026.0697.52.5
MLEvaluation Protocol=Zero-shot2026.0697.312.69
NeMoGuardEvaluation Protocol=Zero-shot2026.0685.5814.42
ProtectAI v1Evaluation Protocol=Zero-shot2026.060.5899.42