Wu, Z., Gao, H., Luo, J., & Liu, Z. (2026). HumorReject: Decoupling LLM Safety from Refusal Prefix via a Little Humor. Proceedings of the AAAI Conference on Artificial Intelligence, 40(44), 38030–38038. https://doi.org/10.1609/aaai.v40i44.41140