@inproceedings{zhang-etal-2025-unlocking,
    title = "Unlocking Recursive Thinking of {LLM}s: Alignment via Refinement",
    author = "Zhang, Haoke  and
      Liang, Xiaobo  and
      Wang, Cunxiang  and
      Li, Juntao  and
      Zhang, Min",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/transition-to-people-yaml/2025.findings-acl.582/",
    doi = "10.18653/v1/2025.findings-acl.582",
    pages = "11169--11182",
    ISBN = "979-8-89176-256-5",
    abstract = "The OpenAI o1-series models have demonstrated that leveraging long-form Chain of Thought (CoT) can substantially enhance performance. However, the recursive thinking capabilities of Large Language Models (LLMs) remain limited, particularly in the absence of expert-curated data for distillation. In this paper, we propose \textbf{AvR}: \textbf{Alignment via Refinement}, a novel method aimed at unlocking the potential of LLMs for recursive reasoning through long-form CoT. AvR introduces a refinement process that integrates criticism and improvement actions, guided by differentiable learning techniques to optimize \textbf{refinement-aware rewards}. As a result, the synthesized multi-round data can be organized as a long refinement thought, further enabling test-time scaling. Experimental results show that AvR significantly outperforms conventional preference optimization methods. Notably, with only 3k synthetic samples, our method boosts the performance of the LLaMA-3-8B-Instruct model by over 20{\%} in win rate on AlpacaEval 2.0. Our code is available at Github ."
}