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

Harmfulness Classification on HarmBench (val)

82.9Agreement Rate

HarmClassifier

73.64476.04778.4580.853Sep 29, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.09
82.9
2025.09
82.5
2025.09
80.6
2025.09
74