@inproceedings{barnhart-etal-2025-aligning, title = "Aligning to What? Limits to {RLHF} Based Alignment", author = "Barnhart, Logan and Akbarian Bafghi, Reza and Becker, Stephen and Raissi, Maziar", editor = "Chiruzzo, Luis and Ritter, Alan and Wang, Lu", booktitle = "Findings of the Association for Computational Linguistics: NAACL 2025", month = apr, year = "2025", address = "Albuquerque, New Mexico", publisher = "Association for Computational Linguistics", url = "https://preview.aclanthology.org/fix-sig-urls/2025.findings-naacl.421/", pages = "7556--7591", ISBN = "979-8-89176-195-7" }