Elamur Research⌕
☰
home / papers / 2026 / GCIG: GraphRAG-based Cross-document Instruction Generation for Boosting LLM Reasoning

GCIG: GraphRAG-based Cross-document Instruction Generation for Boosting LLM Reasoning

Paper · 2026 · Xiaoliang Xu, Huang Yuan, Junmei Wang et al.

Summary

Automatically synthesized instructions are a cheap way to fine-tune LLMs. However, popular generators such as Self-Instruct, WizardLM and Bonito have three weaknesses: they need good seed data, they treat each text chunk in isolation, and they tend to yield repetitive, shallow items that rarely require combining facts from several documents. GCIG tackles this by turning a domain corpus into an enhanced knowledge graph in which every entity and relation stays tied to its source sentences. Around core entities, it extracts K-hop subgraphs via beam search. An LLM then scores the paired text chunks for factual consistency and logical complementarity and discards noisy evidence. Task-aware prompts create varied multi-hop questions, and answers are produced with Chain-of-Thought reasoning that follows entity paths through the graph. Models fine-tuned on GCIG data beat those trained with competing generators on knowledge-intensive and multi-hop QA. Evaluators also preferred its instructions over EntiGraph's. The approach is useful for building reasoning-oriented training sets in specialized domains where labeled seeds are scarce.

Key points

  • The pipeline has four stages: Enhanced KG construction (chunks ≤ 2048 tokens; ⟨entity, relation, context⟩ triples), LLM-driven Chunk Selection (LCS), Adaptive Prompt-based Question Generation (APQG), and CoT-guided Answer Generation (CoTAG).
  • Against EntiGraph, GCIG instructions win 78.1% (Contract) and 83.2% (HotpotQA) of pairwise comparisons on a quality criterion, and 74.6% / 79.1% on information coverage.
  • In the HotpotQA ablation with Mistral, accuracy is 22.7 for the full system, 20.3 without the evidence filter, and 21.6 without CoT. With Qwen2.5 the full system reaches 24.2 and the filter is again the most important component.
  • Stated limitations: building the graph is computationally costly, the pipeline is one-way with no feedback from generation back to the graph, and adding more documents introduces distracting chunks.
  • Code and data are available at github.com/WhitEiller/GCIG.

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

🎓Hangzhou Dianzi UniversityHangzhou · China15 works · 1,061 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksComputer Vision and Pattern Recognition · 3,486 worksInformation Systems · 1,730 works

🧭 Topics 2Research subjects and languages the work is about.Language models & pretraining · 3,894 worksReasoning · 1,685 works

Cite

Cite as Xiaoliang Xu, Huang Yuan, Junmei Wang and Can Xu. 2026. GCIG: GraphRAG-based Cross-document Instruction Generation for Boosting LLM Reasoning. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.871

@inproceedings{xu2026gcig,
  title = {{GCIG: GraphRAG-based Cross-document Instruction Generation for Boosting LLM Reasoning}},
  author = {Xiaoliang Xu, Huang Yuan, Junmei Wang and Can Xu},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.871},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.871}
}

References

44 references, 25 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Akari Asai, Zeqiu Wu, Yizhong Wang, Avirup Sil, and Hannaneh Hajishirzi. 2024. Self-rag: Learning to retrieve, generate, and critique through self-reflection.
  2. Qijie Chen, Haotong Sun, Haoyang Liu, Yinghui Jiang, Ting Ran, Xurui Jin, Xianglu Xiao, Zhimin Lin, Hongming Chen, and Zhangmin Niu. 2023. An extensive benchmark study on biomedical text generation and mining with chatgpt. Bioinformatics, 39(9):btad557. find
  3. Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Yunxuan Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, and 1 others. 2024. Scaling instruction-finetuned language models. Journal of Machine Learning Research, 25(70):1–53. find
  4. Albert Q Jiang, Alexandre Sablayrolles, Antoine Roux, Arthur Mensch, Blanche Savary, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Emma Bou Hanna, Florian Bressand, and 1 others. 2024. Mixtral of experts. arXiv preprint arXiv:2401.04088. find
  5. Tri Dao. 2023. Flashattention-2: Faster attention with better parallelism and work partitioning. arXiv preprint arXiv:2307.08691.
  6. Yuta Koreeda and Christopher Manning. 2021. ContractNLI: A dataset for document-level natural language inference for contracts. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 1907–1919, Punta Cana, Dominican Republic. Association for Computational Linguistics.
  7. Ning Ding, Yulin Chen, Bokai Xu, Yujia Qin, Shengding Hu, Zhiyuan Liu, Maosong Sun, and Bowen Zhou. 2023. Enhancing chat language models by scaling high-quality instructional conversations. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 3029–3051.
  8. Darren Edge, Ha Trinh, Newman Cheng, Joshua Bradley, Alex Chao, Apurva Mody, Steven Truitt, Dasha Metropolitansky, Robert Osazuwa Ness, and Jonathan Larson. 2024. From local to global: A graph rag approach to query-focused summarization. arXiv preprint arXiv:2404.16130. find
  9. Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jinliu Pan, Yuxi Bi, Yixin Dai, Jiawei Sun, Haofen Wang, and Haofen Wang. 2023. Retrieval-augmented generation for large language models: A survey. arXiv preprint arXiv:2312.10997, 2(1).
  10. Woosuk Kwon, Zhuohan Li, Siyuan Zhuang, Ying Sheng, Lianmin Zheng, Cody Hao Yu, Joseph Gonzalez, Hao Zhang, and Ion Stoica. 2023. Efficient memory management for large language model serving with pagedattention. In Proceedings of the 29th symposium on operating systems principles, pages 611–626.
  11. Ming Li, Yong Zhang, Zhitao Li, Jiuhai Chen, Lichang Chen, Ning Cheng, Jianzong Wang, Tianyi Zhou, and Jing Xiao. 2024. From quantity to quality: Boosting llm performance with self-guided data selection for instruction tuning. In Proceedings of the 2024 Conference of the North American Chapter of th…
  12. Arnav Gudibande, Eric Wallace, Charlie Snell, Xinyang Geng, Hao Liu, Pieter Abbeel, Sergey Levine, and Dawn Song. 2023. The false promise of imitating proprietary llms. arXiv preprint arXiv:2305.15717.
  13. Aixin Liu, Bei Feng, Bing Xue, Bingxuan Wang, Bochao Wu, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chenyu Zhang, Chong Ruan, and 1 others. 2024. Deepseek-v3 technical report. arXiv preprint arXiv:2412.19437. find
  14. Neel Guha, Julian Nyarko, Daniel Ho, Christopher Ré, Adam Chilton, Alex Chohlas-Wood, Austin Peters, Brandon Waldon, Daniel Rockmore, Diego Zambrano, and 1 others. 2023. Legalbench: A collaboratively built benchmark for measuring legal reasoning in large language models. Advances in neural informati…
  15. Wanlong Liu, Junying Chen, Ke Ji, Li Zhou, Wenyu Chen, and Benyou Wang. 2025. Rag-instruct: Boosting llms with diverse retrieval-augmented instructions. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 3865–3888.
  16. Suchin Gururangan, Ana Marasović, Swabha Swayamdipta, Kyle Lo, Iz Beltagy, Doug Downey, and Noah A. Smith. 2020. Don’t stop pretraining: Adapt language models to domains and tasks. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 8342–8360, Online. A… find
  17. Or Honovich, Thomas Scialom, Omer Levy, and Timo Schick. 2023. Unnatural instructions: Tuning language models with (almost) no human labor. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 14409–14428.
  18. Hui Huang, Bing Xu, Xinnian Liang, Kehai Chen, Muyun Yang, Tiejun Zhao, and Conghui Zhu. 2024. Multi-view fusion for instruction mining of large language model. Information Fusion, 110:102480. find
  19. Subhabrata Mukherjee, Arindam Mitra, Ganesh Jawahar, Sahaj Agarwal, Hamid Palangi, and Ahmed Awadallah. 2023. Orca: Progressive learning from complex explanation traces of gpt-4. arXiv preprint arXiv:2306.02707.
  20. Nihal Nayak, Yiyang Nan, Avi Trost, and Stephen Bach. 2024. Learning to generate instruction tuning datasets for zero-shot task adaptation. In Findings of the Association for Computational Linguistics: ACL 2024, pages 12585–12611. find
  21. Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, and 1 others. 2022. Training language models to follow instructions with human feedback. Advances in neural information processing systems, 35:27730–27744.
  22. Abhilasha Ravichander, Alan W Black, Shomir Wilson, Thomas Norton, and Norman Sadeh. 2019. Question answering for privacy policies: Combining computational and legal perspectives. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Join…
  23. Victor Sanh, Albert Webson, Colin Raffel, Stephen Bach, Lintang Sutawika, Zaid Alyafeai, Antoine Chaffin, Arnaud Stiegler, Arun Raja, Manan Dey, M Saiful Bari, Canwen Xu, Urmish Thakker, Shanya Sharma Sharma, Eliza Szczechla, Taewoon Kim, Gunjan Chhablani, Nihal Nayak, Debajyoti Datta, and 21 others… — Multitask Prompted Training Enables Zero-Shot Task Generalization
  24. Thomas Savage, Stephen P Ma, Abdessalem Boukil, Ekanath Rangan, Vishwesh Patel, Ivan Lopez, and Jonathan Chen. 2025. Fine-tuning methods for large language models in clinical medicine by supervised fine-tuning and direct preference optimization: Comparative evaluation. Journal of Medical Internet Re… find
  25. Christopher Sciavolino, Zexuan Zhong, Jinhyuk Lee, and Danqi Chen. 2021. Simple entity-centric questions challenge dense retrievers. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 6138–6148, Online and Punta Cana, Dominican Republic. Association for…
  26. Ilia Shumailov, Zakhar Shumaylov, Yiren Zhao, Yarin Gal, Nicolas Papernot, and Ross Anderson. 2023. The curse of recursion: Training on generated data makes models forget. arXiv preprint arXiv:2305.17493.
  27. Zirui Song, Bin Yan, Yuhan Liu, Miao Fang, Mingzhe Li, Rui Yan, and Xiuying Chen. 2025. Injecting domain-specific knowledge into large language models: a comprehensive survey. arXiv preprint arXiv:2502.10708. find
  28. Rohan Taori, Ishaan Gulrajani, Tianyi Zhang, Yann Dubois, Xuechen Li, Carlos Guestrin, Percy Liang, and Tatsunori B Hashimoto. 2023. Alpaca: A strong, replicable instruction-following model. Stanford Center for Research on Foundation Models. https://crfm.stanford. edu/2023/03/13/alpaca. html, 3(6):… ↗
  29. Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, and 1 others. 2023. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288.
  30. Vincent A Traag, Ludo Waltman, and Nees Jan Van Eck. 2019. From louvain to leiden: guaranteeing well-connected communities. Scientific reports, 9(1):1– 12. find
  31. George Tsatsaronis, Georgios Balikas, Prodromos Malakasiotis, Ioannis Partalas, Matthias Zschunke, Michael R Alvers, Dirk Weissenborn, Anastasia Krithara, Sergios Petridis, Dimitris Polychronopoulos, and 1 others. 2015. An overview of the bioasq large-scale biomedical semantic indexing and question… find
  32. Yizhong Wang, Yeganeh Kordi, Swaroop Mishra, Alisa Liu, Noah A Smith, Daniel Khashabi, and Hannaneh Hajishirzi. 2023. Self-instruct: Aligning language models with self-generated instructions. In Proceedings of the 61st annual meeting of the association for computational linguistics (volume 1: long p… find
  33. Zeyuan Wang, Qiang Zhang, Keyan Ding, Ming Qin, Xiang Zhuang, Xiaotong Li, and Huajun Chen. 2024. Instructprotein: Aligning human and protein language via knowledge instruction. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages… find
  34. Jason Wei, Maarten Bosma, Vincent Y Zhao, Kelvin Guu, Adams Wei Yu, Brian Lester, Nan Du, Andrew M Dai, and Quoc V Le. 2021. Finetuned language models are zero-shot learners. arXiv preprint arXiv:2109.01652.
  35. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, and 1 others. 2022. Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35:24824– 24837.
  36. Can Xu, Qingfeng Sun, Kai Zheng, Xiubo Geng, Pu Zhao, Jiazhan Feng, Chongyang Tao, Qingwei Lin, and Daxin Jiang. 2024. Wizardlm: Empowering large pre-trained language models to follow complex instructions. In The Twelfth International Conference on Learning Representations.
  37. An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, and 1 others. 2025a. Qwen2.5 technical report. arXiv preprint, arXiv:2412.15115. find
  38. Zhilin Yang, Peng Qi, Saizheng Zhang, Yoshua Bengio, William Cohen, Ruslan Salakhutdinov, and Christopher D Manning. 2018. Hotpotqa: A dataset for diverse, explainable multi-hop question answering. In Proceedings of the 2018 conference on empirical methods in natural language processing, pages 2369–…
  39. Zitong Yang, Neil Band, Shuangping Li, Emmanuel Candes, and Tatsunori Hashimoto. 2025b. Synthetic continued pretraining. In International Conference on Representation Learning, volume 2025, pages 44379–44421. find
  40. Shengyu Zhang, Linfeng Dong, Xiaoya Li, Sen Zhang, Xiaofei Sun, Shuhe Wang, Jiwei Li, Runyi Hu, Tianwei Zhang, Guoyin Wang, and 1 others. 2023. Instruction tuning for large language models: A survey. ACM Computing Surveys.
  41. Hyperparameters Values Instruction Generation temperature 1.5 top_p 1.0 presence_penalty 0.3 frequency_penalty 0.3 max_tokens 16384 find
  42. Xuanyu Zhang and Qing Yang. 2023. Self-qa: Unsupervised knowledge guided language model alignment. arXiv preprint arXiv:2305.11952.
  43. Response Generation temperature 0.3 top_p 1.0 presence_penalty 0.5 frequency_penalty 0.5 max_tokens 16384 find
  44. Yaowei Zheng, Richong Zhang, Junhao Zhang, Yanhan Ye, Zheyan Luo, Zhangchi Feng, and Yongqiang Ma. 2024. Llamafactory: Unified efficient fine-tuning of 100+ language models. arXiv preprint arXiv:2403.13372.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/xu-2026-gcig-graphrag-based-cross-document/