Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

HarmBench and DEFCON

Benchmarks

Task NameDataset NameSOTA ResultTrend
Multi-turn Jailbreak DetectionHarmBench and DEFCON Multi-turn Jailbreak N=1,010 (test)
F1 Score84
10
Showing 1 of 1 rows