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

Content Safety Detection on HarmBench

95.6F1 Score

prompt–response verification framework

90.08891.51992.9594.381Jun 24, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
95.63.9
2026.06
90.3-