List your company

Jailbreak Detection on Red-teaming Queries high-risk vulnerabilities

1.871High-Risk Vulnerability Discovery RateLast Generation
0.450360.819181.1881.55682Jan 8, 2026

Evaluation results

MethodLinks
Last GenerationRelative Compute Overhead (%)=100.02026.011.871
Exchange ClassifierRelative Compute Overhead (%)=150.02026.010.885
Two-Stage CascadeRelative Compute Overhead (%)=27.82026.010.878
Production GradeRelative Compute Overhead (%)=3.52026.010.505