Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

DeepInception

Benchmarks

Task NameDataset NameSOTA ResultTrend
Jailbreak DefenseDeepInception
Harmful Score1
58
Safety EvaluationDeepInception (test)
ASR78.5
7
Safety MonitoringDeepInception
ASR @ 1% FPR67.02
5
Showing 3 of 3 rows