@inproceedings{chen-etal-2025-extracting,
    title = "Extracting and Understanding the Superficial Knowledge in Alignment",
    author = "Chen, Runjin  and
      Perin, Gabriel Jacob  and
      Chen, Xuxi  and
      Chen, Xilun  and
      Han, Yan  and
      Hirata, Nina S. T.  and
      Hong, Junyuan  and
      Kailkhura, Bhavya",
    editor = "Chiruzzo, Luis  and
      Ritter, Alan  and
      Wang, Lu",
    booktitle = "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)",
    month = apr,
    year = "2025",
    address = "Albuquerque, New Mexico",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/fix-sig-urls/2025.naacl-long.168/",
    pages = "3265--3280",
    ISBN = "979-8-89176-189-6"
}