Search papers, labs, and topics across Lattice.
This paper introduces CheckRLM, a novel framework that enhances the reliability of Reasoning Language Models (RLMs) by integrating Retrieval-Augmented Generation (RAG) to check and correct factual errors in reasoning chains. By extracting factual claims during inference, CheckRLM identifies and localizes knowledge inconsistencies, applying a refinement mechanism that utilizes external knowledge for precise corrections. Experimental results show that CheckRLM significantly reduces error accumulation in long-horizon reasoning tasks while maintaining lower operational costs compared to existing methods.
CheckRLM cuts error accumulation in reasoning chains by correcting factual inaccuracies in real-time, outperforming traditional approaches.
Reasoning Language Models (RLMs) have significantly improved performance on complex tasks by extending the reasoning chain. However, these chains are prone to containing factual errors, particularly in knowledge-intensive tasks. To address this issue, we propose CheckRLM, a framework that improves the reliability of the reasoning process through Retrieval-Augmented Generation (RAG) by timely checking and correcting factual errors. Specifically, CheckRLM extracts factual claims from the reasoning chain to identify and localize subtle knowledge inconsistencies during inference. Upon detection of errors, a refinement mechanism performs minimal-cost yet precise corrections by leveraging external knowledge, ensuring coherence between the reasoning chain and correct knowledge. Extensive experiments demonstrate that CheckRLM substantially outperforms existing baselines, exhibiting a strong capability to mitigate error accumulation in long-horizon reasoning with lower costs. The code and data are available at https://github.com/AI9Stars/CheckRLM.
Tsinghua AI