arXiv 2508.012732026-07-01cs.AI

解耦推理逻辑以解决显式知识冲突

Disentangling Reasoning Logic to Resolve Explicit Knowledge Conflicts

School of Computer Science, University of Auckland（奥克兰大学计算机科学学院）；Department of Electronics and Electrical Engineering, National Yang Ming Chiao Tung University（国立阳明交通大学电子与电机工程学系）；School of Computing and Information Systems, Singapore Management University（新加坡管理大学计算与信息系统学院）；Shenzhen Key Laboratory of Internet Information Collaboration, Harbin Institute of Technology (Shenzhen)（哈尔滨工业大学（深圳）深圳市互联网信息协同重点实验室）

Xianda Zheng, Zijian Huang, Meng-Fen Chiang, Jiamou Liu, Yuan Fang, Michael Witbrock, Kaiqi Zhao

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出KCR框架，通过将冲突上下文解耦为推理轨迹并利用强化学习优化逻辑一致性，有效解决LLM中的显式知识冲突，7B模型性能超越GPT-4o等。

摘要：

显式知识冲突发生在检索到的上下文包含矛盾信息时，随着大语言模型（LLMs）整合日益多样化的数据源，这一挑战变得根本。核心困难在于纠缠叙述和异质冲突模式的复杂性，这常常超出标准骨干架构的推理能力。我们提出\textbf{\textsc{Kcr}}（知识冲突推理）框架，通过系统性地构建其底层逻辑来裁决矛盾。\textsc{Kcr}将冲突上下文解耦为离散的推理轨迹集，利用文本和图的混合表示促进系统理解。然后，它采用可验证奖励的强化学习（RLVR）范式来灌输推理策略，最大化逻辑一致性，同时抑制源自矛盾证据的虚假路径。广泛评估表明，\textsc{Kcr}带来了显著的性能提升。值得注意的是，由\textsc{Kcr}增强的7B模型在复杂任务上的裁决能力显著优于领先的专有模型，包括GPT-4o和GPT-5.1。代码见此https URL。

英文摘要：

Explicit knowledge conflicts, occurring when retrieved contexts contain contradictory information, pose a fundamental challenge for Large Language Models (LLMs) as they integrate increasingly diverse data sources. The core difficulty lies in the complexity of entangled narratives and heterogeneous conflict patterns, which frequently exceeds the reasoning capacity of standard backbone architectures. We propose \textbf{\textsc{Kcr}} (Knowledge Conflict Reasoning), a framework that adjudicates contradictions by systematically structuring their underlying logic. \textsc{Kcr} disentangles conflicting contexts into discrete sets of reasoning traces, utilizing a hybrid representation of text and graphs to facilitate systematic comprehension. It then employs a Reinforcement Learning with Verifiable Rewards (RLVR) paradigm to instill a reasoning policy that maximizes logical consistency while suppressing spurious paths derived from contradictory evidence. Extensive evaluations demonstrate that \textsc{Kcr} yields substantial performance gains. Notably, a 7B model enhanced by \textsc{Kcr} achieves adjudication capabilities that significantly outperform leading proprietary models, including GPT-4o and GPT-5.1, on complex tasks. Code is available at https://github.com/zhengxianda/KCR.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。