@inproceedings{jiang-etal-2026-mitigating,
    title = "Mitigating Over-Refusal in Aligned Large Language Models via Inference-Time Activation Energy",
    author = "Jiang, Eric Hanchen  and
      Ou, Weixuan  and
      Liu, Run  and
      Pang, Shengyuan  and
      Wan, Guancheng  and
      Duan, Ranjie  and
      Dong, Wei  and
      Chang, Kai-Wei  and
      Wang, XiaoFeng  and
      Wu, Ying Nian  and
      Li, Xinfeng",
    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.1759/",
    pages = "37930--37950",
    ISBN = "979-8-89176-390-6",
    abstract = "Safety alignment of large language models currently faces a central challenge: existing alignment techniques often prioritize mitigating responses to harmful prompts at the expense of overcautious behavior, leading models to incorrectly refuse benign requests. A key goal of safe alignment is therefore to improve safety while simultaneously minimizing false refusals. In this work, we introduce Energy Landscape Steering (ELS), a novel, fine-tuning free framework designed to resolve this challenge through dynamic, inference-time intervention. We trained a lightweight, external Energy-Based Model (EBM) to assign high energy to undesirable (false refusal or jailbreak) states and low energy to desirable (helpful response or safe reject) ones. During inference, the EBM maps the LLM{'}s internal activations to an energy landscape, and we use the gradient of the energy function to steer the hidden states toward low-energy regions in real time. This dynamically guides the model toward desirable behavior without modifying its parameters. By decoupling behavioral control from the model{'}s core knowledge, ELS provides a flexible and computationally efficient solution. Extensive experiments across diverse models demonstrate its effectiveness: raising compliance on the ORB-H benchmark from 57.3{\%} to 82.6{\%} while maintaining the baseline safety performance. Our work establishes a promising paradigm for building LLMs that simultaneously achieve high safety and low false refusal rates."
}