ResearchBenchmarksJailbreaking on NSFW-200 1.0 (test)Follow90.5Bypass RateOptJail58.2666.637583.37May 25, 2025Evaluation ResultsMethodMethodLinksBypass RateQuery MetricOptJailSafety Mechanism=SLDSafety Mechanism=SLD2025.0590.56.8OptJailSafety Mechanism=UCESafety Mechanism=UCE2025.0582.52.9OptJailSafety Mechanism=SafeGenSafety Mechanism=SafeGen2025.0559.512.95