Zhang, Xiaoyun, Zhengyue Zhao, Wenxuan Shi, Kaidi Xu, Di Huang, and Xing Hu. 2026. “Safety Alignment of Large Language Models via Contrasting Safe and Harmful Distributions”. Proceedings of the AAAI Conference on Artificial Intelligence 40 (41):34827-35. https://doi.org/10.1609/aaai.v40i41.40785.