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

Unsafe Prompts

Benchmarks

Task NameDataset NameSOTA ResultTrend
JailbreakingUnsafe Prompts
Bypass Success Rate (Text)98.5
22
Safety DefenseUnsafe Prompts 7 categories
Hate Rate100
4
Showing 2 of 2 rows