MR-ALIGN: Meta-Reasoning Informed Factuality Alignment for Large Reasoning Models

ACL ARR 2026 January Submission1713 Authors

Keywords: Factuality Alingment, Reasoning, Factual QA

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 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 and one long-form factuality benchmarks 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.

Paper Type: Long

Research Area: Dialogue and Interactive Systems

Research Area Keywords: factuality,reasoning,fine-tuning

Contribution Types: NLP engineering experiment

Languages Studied: English

Submission Number: 1713