@inproceedings{xu-etal-2026-steering,
    title = "Why Steering Works: Toward a Unified View of Language Model Parameter Dynamics",
    author = "Xu, Ziwen  and
      WU, Chenyan  and
      Sun, Hengyu  and
      Hong, Haiwen  and
      Wang, Mengru  and
      Yao, Yunzhi  and
      Huang, Longtao  and
      Xue, Hui  and
      Deng, Shumin  and
      Chu, Zhixuan  and
      Chen, Huajun  and
      Zhang, Ningyu",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.1463/",
    pages = "31719--31736",
    ISBN = "979-8-89176-390-6",
    abstract = "Methods for controlling large language models (LLMs), including local weight fine-tuning, LoRA-based adaptation, and activation-based interventions, are often studied in isolation, obscuring their connections and making comparison difficult. In this work, we present a unified view that frames these interventions as dynamic weight updates induced by a control signal, placing them within a single conceptual framework. Building on this view, we propose a unified preference-utility analysis that separates control effects into preference, defined as the tendency toward a target concept, and utility, defined as coherent and task-valid generation, and measures both on a shared log-odds scale using polarity-paired contrastive examples. Across methods, we observe a consistent trade-off between preference and utility: stronger control increases preference while predictably reducing utility. We further explain this behavior through an activation manifold perspective, in which control shifts representations along target-concept directions to enhance preference, while utility declines primarily when interventions push representations off the model{'}s valid-generation manifold. Finally, we introduce a new steering approach guided by this analysis that improves preference while better preserving utility."
}