@inproceedings{tang-etal-2025-unlocking,
    title = "Unlocking General Long Chain-of-Thought Reasoning Capabilities of Large Language Models via Representation Engineering",
    author = "Tang, Xinyu  and
      Wang, Xiaolei  and
      Lv, Zhihao  and
      Min, Yingqian  and
      Zhao, Xin  and
      Hu, Binbin  and
      Liu, Ziqi  and
      Zhang, Zhiqiang",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingestion-acl-25/2025.acl-long.339/",
    pages = "6832--6849",
    ISBN = "979-8-89176-251-0",
    abstract = "Recent advancements in long chain-of-thoughts (long CoTs) have significantly improved the reasoning capabilities of large language models (LLMs). Existing work finds that the capability of long CoT reasoning can be efficiently elicited by tuning on only a few examples and can easily transfer to other tasks. This motivates us to investigate whether long CoT reasoning is a general capability for LLMs. In this work, we conduct an empirical analysis for this question from the perspective of representation. We find that LLMs do encode long CoT reasoning as a general capability, with a clear distinction from vanilla CoTs. Furthermore, domain-specific representations are also required for the effective transfer of long CoT reasoning. Inspired by these findings, we propose GLORE, a novel representation engineering method to unleash the general long CoT reasoning capabilities of LLMs. Extensive experiments demonstrate the effectiveness and efficiency of GLORE in both in-domain and cross-domain scenarios. The code is available at https://github.com/txy77/GLoRE."
}