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

PKU-Safe

Benchmarks

Task NameDataset NameSOTA ResultTrend
Harmful QueryPKU-Safe
ASR0.52
20
LLM Safety and Informativeness EvaluationPKU-Safe
Safety Rate90.83
11
Adversarial Robustness EvaluationPKU-Safe
Attack Success Rate (ASR)25
4
Showing 3 of 3 rows