@inproceedings{zhong-etal-2025-react,
    title = "{REACT}: Representation Extraction And Controllable Tuning to Overcome Overfitting in {LLM} Knowledge Editing",
    author = "Zhong, Haitian  and
      Liu, Yuhuan  and
      Xu, Ziyang  and
      Liu, Guofan  and
      Liu, Qiang  and
      Wu, Shu  and
      Zhao, Zhe  and
      Wang, Liang  and
      Tan, Tieniu",
    editor = "Christodoulopoulos, Christos  and
      Chakraborty, Tanmoy  and
      Rose, Carolyn  and
      Peng, Violet",
    booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-emnlp/2025.emnlp-main.860/",
    pages = "16994--17011",
    ISBN = "979-8-89176-332-6",
    abstract = "Large language model editing methods frequently suffer from overfitting, wherein factual updates can propagate beyond their intended scope, overemphasizing the edited target even when it{'}s contextually inappropriate. To address this challenge, we introduce REACT (Representation Extraction And Controllable Tuning), a unified two-phase framework designed for precise and controllable knowledge editing. In the initial phase, we utilize tailored stimuli to extract latent factual representations and apply Principal Component Analysis with a simple learnbale linear transformation to compute a directional ``belief shift'' vector for each instance. In the second phase, we apply controllable perturbations to hidden states using the obtained vector with a magnitude scalar, gated by a pre-trained classifier that permits edits only when contextually necessary. Relevant experiments on EVOKE benchmarks demonstrate that REACT significantly reduces overfitting across nearly all evaluation metrics, and experiments on COUNTERFACT and MQuAKE shows that our method preserves balanced basic editing performance (reliability, locality, and generality) under diverse editing scenarios."
}