--- title: "Disentangling Reasoning Logic to Resolve Explicit Knowledge" description: "They argue that existing fusion decoders and prompting heuristics fail for two reasons: contradictory evidence gets tangled… 4 references." authors: ["Xianda Zheng", "Zijian Huang", "Meng-Fen Chiang", "Jiamou Liu", "Yuan Fang", "Michael J. Witbrock", "Kaiqi Zhao"] year: 2026 venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)" source: "https://aclanthology.org/2026.acl-long.1451/" doi: "10.18653/v1/2026.acl-long.1451" licence: "CC BY 4.0 (ACL Anthology)" licence_url: "https://creativecommons.org/licenses/by/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/zheng-2026-disentangling-reasoning-logic-resolve-explicit/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/zheng-2026-disentangling-reasoning-logic-resolve-explicit/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/30/) / Disentangling Reasoning Logic to Resolve Explicit Knowledge Conflicts # Disentangling Reasoning Logic to Resolve Explicit Knowledge Conflicts **Authors:** [Xianda Zheng](https://md.research.elamur.ai/free/en/authors/xianda-zheng/), Zijian Huang, [Meng-Fen Chiang](https://md.research.elamur.ai/free/en/authors/meng-fen-chiang/), Jiamou Liu, Yuan Fang, [Michael J. Witbrock](https://md.research.elamur.ai/free/en/authors/michael-witbrock/) and [Kaiqi Zhao](https://md.research.elamur.ai/free/en/authors/kaiqi-zhao/) · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2026.acl-long.1451/) · [PDF](https://aclanthology.org/2026.acl-long.1451.pdf) · [DOI 10.18653/v1/2026.acl-long.1451](https://doi.org/10.18653/v1/2026.acl-long.1451) · [OpenAlex](https://openalex.org/W7166879923) **Citations:** 0 citations (OpenAlex, 2026-10-14) ## Summary When retrieved documents give opposing answers to the same question, an LLM has to decide which source to trust. The authors call these explicit knowledge conflicts. They argue that existing fusion decoders and prompting heuristics fail for two reasons: contradictory evidence gets tangled together (logic entanglement), and conflict types vary too much for fixed rules (case heterogeneity). Their framework, KCR (Knowledge Conflict Reasoning), works in two stages. First, it extracts the key entity and relation from the query and then rewrites each candidate answer's context into separate reasoning traces. These traces come in two forms: textual entity–relation chains and paths through a local knowledge graph. Second, it trains the backbone model with RLVR using two rewards. A logic coherence reward pulls the model toward the traces of the correct answer and away from the spurious ones. A logic consistency reward penalizes mixing mutually exclusive facts. The task follows the ConflictQA setup. With KCR, a 7B model beats frameworks built on GPT-4o and GPT-5.1. This suggests that structured reasoning training, rather than model scale, can drive reliable conflict adjudication in RAG. ## Key points - **Task setup:** each input has a query, two conflicting candidate answers, and a supporting context for each, following ConflictQA (Xie et al., 2024). - **Disentanglement stage:** - A frozen Qwen2.5-7B does zero-shot entity and relation extraction, with popQA entity labels used where available. - GPT-4o-mini builds the local knowledge graphs. - Only graph paths that contain the query entity or relation are kept. - **Training stage:** RLVR with group sampling and logic coherence plus logic consistency rewards. The authors chose RL over SFT because SFT tends to memorize specific conflict patterns. - **Main results:** on the Inter-Context Knowledge Conflict task, the best KCR variant beats the strongest prior framework by between +2.4% and +24.3% in the reported comparison. All results are averaged over three runs. - **Ablations and limitations:** - Both rewards turn out to be necessary. - Gains are largest on tasks with strict logical chains, such as popQA. - The method depends on external APIs for graph extraction and costs more to train than SFT. - Evaluation covers only controlled, non-adversarial settings. - Code is released at github.com/zhengxianda/KCR. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* ## Institutions - [Harbin Institute of Technology](https://md.research.elamur.ai/free/en/universities/harbin-institute-of-technology/) — University, Harbin, China · 185 works · 29,140 citations - [Singapore Management University](https://md.research.elamur.ai/free/en/universities/singapore-management-university/) — University, Singapore, Singapore · 35 works · 2,457 citations - [University of Auckland](https://md.research.elamur.ai/free/en/universities/university-of-auckland/) — University, Auckland, New Zealand · 15 works · 888 citations - [National Yang Ming Chiao Tung University](https://md.research.elamur.ai/free/en/universities/national-yang-ming-chiao-tung-university/) — University, Hsinchu, Taiwan · 7 works · 85 citations 📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works 🧭 Topics (1): [🧮 Reasoning](https://md.research.elamur.ai/free/en/topics/reasoning/) · 1,685 works ## Cite Cite as: Xianda Zheng, Zijian Huang, Meng-Fen Chiang, Jiamou Liu, Yuan Fang, Michael J. Witbrock and Kaiqi Zhao. 2026. Disentangling Reasoning Logic to Resolve Explicit Knowledge Conflicts. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{zheng2026disentangling, title = {{Disentangling Reasoning Logic to Resolve Explicit Knowledge Conflicts}}, author = {Xianda Zheng, Zijian Huang, Meng-Fen Chiang, Jiamou Liu, Yuan Fang, Michael J. Witbrock and Kaiqi Zhao}, booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)}, year = {2026}, doi = {10.18653/v1/2026.acl-long.1451}, url = {https://doi.org/10.18653/v1/2026.acl-long.1451} } ``` - [← previous Disentangling Linguistic Relatedness from Task Alignment in Cross-Lingual Transfer](https://md.research.elamur.ai/free/en/papers/ahmed-2026-disentangling-linguistic-relatedness-task-alignment/) - [next → Disentangling the Effects of Unlearning in Measuring Parametric Faithfulness of Chain-of-Thought](https://md.research.elamur.ai/free/en/papers/mitsuhashi-2026-disentangling-effects-unlearning-measuring-parametric/) - [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/30/) ## References 4 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Jinze Bai, Shuai Bai, Yunfei Chu, Zeyu Cui, Kai Dang, Xiaodong Deng, Yang Fan, Wenbin Ge, Yu Han, Fei Huang, Binyuan Hui, Luo Ji, Mei Li, Junyang Lin, Runji Lin, Dayiheng Liu, Gao Liu, Chengqiang Lu, Keming Lu, and 29 others. 2023. [Qwen technical report](https://md.research.elamur.ai/free/en/papers/bai-2023-qwen-technical-report/). CoRR, abs/2309.16609. 2. Darren Edge, Ha Trinh, Newman Cheng, Joshua Bradley, Alex Chao, Apurva Mody, Steven Truitt, and Jonathan Larson. 2024. From local to global: A graph RAG approach to query-focused summarization. CoRR, abs/2404.16130. [find](https://scholar.google.com/scholar?q=Darren+Edge%2C+Ha+Trinh%2C+Newman+Cheng%2C+Joshua+Bradley%2C+Alex+Chao%2C+Apurva+Mody%2C+Steven+Truitt%2C+and+Jonathan+Larson%2E+2024%2E+From+local+to+global%3A+A+graph+RAG+approach+to+query%2Dfocused+summarization%2E+CoRR%2C+) 3. Rishin Haldar and Debajyoti Mukhopadhyay. 2011. Levenshtein distance technique in dictionary lookup methods: An improved approach. CoRR, abs/1101.1232. — [arXiv:1101.1232](https://arxiv.org/abs/1101.1232) [find](https://scholar.google.com/scholar?q=Rishin+Haldar+and+Debajyoti+Mukhopadhyay%2E+2011%2E+Levenshtein+distance+technique+in+dictionary+lookup+methods%3A+An+improved+approach%2E+CoRR%2C+abs%2F1101%2E1232%2E+%E2%80%94+%5BarXiv%3A1101%2E1232%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F1101%2E1) 4. Giwon Hong, Jeonghwan Kim, Junmo Kang, Sung-Hyon Myaeng, and Joyce Jiyoung Whang. 2024. Why so gullible? enhancing the robustness of retrieval-augmented models against counterfactual noise. In NAACL Findings, pages 2474–2495. [find](https://scholar.google.com/scholar?q=Giwon+Hong%2C+Jeonghwan+Kim%2C+Junmo+Kang%2C+Sung%2DHyon+Myaeng%2C+and+Joyce+Jiyoung+Whang%2E+2024%2E+Why+so+gullible%3F+enhancing+the+robustness+of+retrieval%2Daugmented+models+against+counterfactual+noise%2E+In+NAACL+F) *Updated 2 November 2026*