Xuanyi Fu, Danbing Zou. A Comparative Empirical Study of Over-Refusal Behavior in Closed-Source Large Language Models on Pseudo-Harmful Prompts. JAIR [Internet]. 2025 Jul. 10 [cited 2026 Oct. 10];6(3):34-45. Available from: https://learnedvertex.com/index.php/JAIR/article/view/361