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

Sorry

Benchmarks

Task NameDataset NameSOTA ResultTrend
JailbreakSorry
Jailbreak Rate99.8
70
Harmful QuerySORRY
ASR (%)12.79
20
Prompt-side safety risk classificationSorry
F1 Score84.6
12
Jailbreak Attack EvaluationSORRY
Attack Success Count28
2
Showing 4 of 4 rows