@inproceedings{wang-etal-2026-mr,
    title = "{MR}-{ALIGN}: Meta-Reasoning Informed Factuality Alignment for Large Reasoning Models",
    author = "Wang, Xinming  and
      Xu, Jian  and
      Yu, Bin  and
      Lian, Sheng  and
      Chen, yi  and
      Wang, Boran  and
      Zhu, Yingjian  and
      Yi, Hongzhu  and
      Yang, Hong-Ming  and
      Hu, Han  and
      Liu, Cheng-Lin  and
      Zhang, Xu-Yao",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.204/",
    pages = "4191--4214",
    ISBN = "979-8-89176-395-1",
    abstract = "Large reasoning models (LRMs) show strong capabilities in complex reasoning, yet their marginal gains on evidence-dependent factual questions are limited. We find this limitation is partially attributable to a reasoning{--}answer hit gap, where the model identifies the correct facts during reasoning but fails to incorporate them into the final response, thereby reducing factual fidelity. To address this issue, we propose MR-ALIGN, a Meta-Reasoning informed alignment framework that enhances factuality without relying on external verifiers. MR-ALIGN quantifies state-transition probabilities along the model{'}s thinking process and constructs a transition-aware implicit reward that reinforces beneficial reasoning patterns while suppressing defective ones at the atomic thinking segments. This re-weighting reshapes token-level signals into probability-aware segment scores, encouraging coherent reasoning trajectories that are more conducive to factual correctness. Empirical evaluations across four factual QA datasets and one long-form factuality benchmark show that MR-ALIGN consistently improves accuracy and truthfulness while reducing misleading reasoning. These results highlight that aligning the reasoning process itself, rather than merely the outputs, is pivotal for advancing factuality in LRMs."
}