Elamur Research⌕
☰
home / papers / 2026 / ACR: Adaptive Context Refactoring via Context Refactoring Operators for Multi-Turn Dialogue

ACR: Adaptive Context Refactoring via Context Refactoring Operators for Multi-Turn Dialogue

Paper · 2026 · Jiawei Shen, Jia Zhu, Hanghui Guo et al.

Summary

Long conversations cause LLMs to lose track of earlier facts and goals. The authors name two failure modes. In contextual inertia, early mistakes keep being reused and reinforced. In state drift, key constraints and intermediate results get buried in noise or lost during compression. Larger context windows, external memory, RAG and compression mostly supply more information but do not correct the history itself.

The authors propose ACR (Adaptive Context Refactoring). An external controller watches the dialogue history and, when needed, rewrites it before the reasoning model sees it. The controller has two parts:

  • a Router, which decides whether and which operator to apply;
  • a Refactorer, which carries out the rewrite.

Both are trained with a teacher-guided self-evolving scheme instead of costly reinforcement learning. Evaluation covers single-hop and multi-hop QA benchmarks. ACR beats retrieval baselines such as DRAGIN and DioR. It scores slightly below RL methods like Search-R1 but uses far fewer tokens. This separates context management from reasoning, a practical option when compute is limited.

Key points

  • The operator library covers six refactoring strategy types, including a NONE option for no intervention. Training moves from a supervised cold start to self-iteration as teacher involvement gradually decreases.
  • ACR generates about 278.54 tokens on average versus about 1678 for Search-R1 (−83.4%). The reductions are −54.2% vs DioR, −37.1% vs DRAGIN and −35.8% vs StepSearch.
  • Benchmarks: NQ, TriviaQA, PopQA, HotpotQA, 2WikiMultiHopQA, MuSiQue and Bamboogle. Full ACR reaches, for example, 35.10 EM on HotpotQA and 36.36 EM on Bamboogle.
  • In ablations, removing the Router hurts most, costing about 13.8 EM on HotpotQA and 2Wiki. Removing the Refactorer costs 10.57 EM on HotpotQA. A prompt-only Qwen2.5-7B-Instruct controller does much worse than the trained one.
  • Limitations: the extra modules add latency and memory. Evaluation is QA-only, with no agentic or tool-use tasks. Code is available at github.com/ClannadKno/multi-turn.

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

🎓Hong Kong University of Science and TechnologyHong Kong · China142 works · 25,136 citations🏢Alibaba GroupHangzhou · China168 works · 15,386 citations🎓Soochow UniversitySuzhou · China63 works · 4,507 citations🎓Southeast UniversityNanjing · China28 works · 354 citations🎓Southeast UniversityDhaka · Bangladesh8 works · 160 citations🎓Zhejiang Normal UniversityJinhua · China4 works · 16 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 1Research subjects and languages the work is about.Dialogue systems · 624 works

Cite

Cite as Jiawei Shen, Jia Zhu, Hanghui Guo, Weijie Shi, Yue Cui, Qingyu Niu, Guoqing Ma, Jingjiang Liu, Yidan Liang, Yilin Wang, Shimin Di and Jiajie Xu. 2026. ACR: Adaptive Context Refactoring via Context Refactoring Operators for Multi-Turn Dialogue. 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.155

@inproceedings{shen2026acr,
  title = {{ACR: Adaptive Context Refactoring via Context Refactoring Operators for Multi-Turn Dialogue}},
  author = {Jiawei Shen, Jia Zhu, Hanghui Guo, Weijie Shi, Yue Cui, Qingyu Niu, Guoqing Ma, Jingjiang Liu, Yidan Liang, Yilin Wang, Shimin Di and Jiajie Xu},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.155},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.155}
}

References

36 references, 20 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Sanghwan Bae, Donghyun Kwak, Soyoung Kang, Min Young Lee, Sungdong Kim, Yuin Jeong, Hyeri Kim, Sang-Woo Lee, Woomyoung Park, and Nako Sung. 2022. Keep me updated! memory management in long-term conversations. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 3769–3787,… find
  2. Yushi Bai, Xin Lv, Jiajie Zhang, Hongchang Lyu, Jiankai Tang, Zhidian Huang, Zhengxiao Du, Xiao Liu, Aohan Zeng, Lei Hou, Yuxiao Dong, Jie Tang, and Juanzi Li. 2024. LongBench: A bilingual, multi-task benchmark for long context understanding. In Proceedings of the 62nd Annual Meeting of the Associat… find
  3. Yukang Chen, Shengju Qian, Haotian Tang, Xin Lai, Zhijian Liu, Song Han, and Jiaya Jia. 2024. Longlora: Efficient fine-tuning of long-context large language models. In The International Conference on Learning Representations (ICLR).
  4. Yu-Neng Chuang, Tianwei Xing, Chia-Yuan Chang, Zirui Liu, Xun Chen, and Xia Hu. 2024. Learning to compress prompt in natural language formats. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1:… find
  5. 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
  6. Florin Cuconasu, Giovanni Trappolini, Federico Siciliano, Simone Filice, Cesare Campagnano, Yoelle Maarek, Nicola Tonellotto, and Fabrizio Silvestri. 2024. The power of noise: Redefining retrieval for rag systems. In Proceedings of the 47th International ACM SIGIR Conference on Research and Developm…
  7. Tri Dao, Daniel Y. Fu, Stefano Ermon, Atri Rudra, and Christopher Ré. 2022. Flashattention: fast and memory-efficient exact attention with io-awareness. In Proceedings of the 36th International Conference on Neural Information Processing Systems, NIPS ’22, Red Hook, NY, USA. Curran Associates Inc. find
  8. Nouha Dziri, Sivan Milton, Mo Yu, Osmar Zaiane, and Siva Reddy. 2022. On the origin of hallucinations in conversational models: Is it the datasets or the models? In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Techn…
  9. Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, and 1 others. 2025a. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948.
  10. Hanghui Guo, Jia Zhu, Shimin Di, Weijie Shi, Zhangze Chen, and Jiajie Xu. 2025b. DioR: Adaptive cognitive detection and contextual retrieval optimization for dynamic retrieval-augmented generation. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1:… find
  11. Xanh Ho, Anh-Khoa Duong Nguyen, Saku Sugawara, and Akiko Aizawa. 2020. Constructing a multi-hop qa dataset for comprehensive evaluation of reasoning steps. In Proceedings of the 28th International Conference on Computational Linguistics, pages 6609– 6625.
  12. Huiqiang Jiang, Qianhui Wu, Chin-Yew Lin, Yuqing Yang, and Lili Qiu. 2023. LLMLingua: Compressing prompts for accelerated inference of large language models. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 13358–13376, Singapore. Association for Compu…
  13. Zhengbao Jiang, Luyu Gao, Zhiruo Wang, Jun Araki, Haibo Ding, Jamie Callan, and Graham Neubig. 2022. Retrieval as attention: End-to-end learning of retrieval and reading within a single transformer. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 2336… find
  14. Bernal Jimenez Gutierrez, Yiheng Shu, Yu Gu, Michihiro Yasunaga, and Yu Su. 2024. Hipporag: Neurobiologically inspired long-term memory for large language models. Advances in Neural Information Processing Systems, 37:59532–59569.
  15. Bowen Jin, Hansi Zeng, Zhenrui Yue, Jinsung Yoon, Sercan Arik, Dong Wang, Hamed Zamani, and Jiawei Han. 2025. Search-r1: Training llms to reason and leverage search engines with reinforcement learning. arXiv preprint arXiv:2503.09516.
  16. Mandar Joshi, Eunsol Choi, Daniel S Weld, and Luke Zettlemoyer. 2017. Triviaqa: A large scale distantly supervised challenge dataset for reading comprehension. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1601–1611.
  17. Tom Kwiatkowski, Jennimaria Palomaki, Olivia Redfield, Michael Collins, Ankur Parikh, Chris Alberti, Danielle Epstein, Illia Polosukhin, Jacob Devlin, Kenton Lee, and 1 others. 2019. Natural questions: a benchmark for question answering research. Transactions of the Association for Computational Lin…
  18. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, and Douwe Kiela. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. In Proceedings of the 34th Internation…
  19. Nelson F. Liu, Kevin Lin, John Hewitt, Ashwin Paranjape, Michele Bevilacqua, Fabio Petroni, and Percy Liang. 2024. Lost in the middle: How language models use long contexts. Transactions of the Association for Computational Linguistics, 12:157–173.
  20. Alex Mallen, Akari Asai, Victor Zhong, Rajarshi Das, Daniel Khashabi, and Hannaneh Hajishirzi. 2023. When not to trust language models: Investigating effectiveness of parametric and non-parametric memories. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (V… find
  21. Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, John Schulman, Jacob Hilton, Fraser Kelton, Luke Miller, Maddie Simens, Amanda Askell, Peter Welinder, Paul F Christiano, Jan Leike, and Ryan Lowe. 2022. Tra… — Training language models to follow instructions with human feedback
  22. Ofir Press, Muru Zhang, Sewon Min, Ludwig Schmidt, Noah A Smith, and Mike Lewis. 2023. Measuring and narrowing the compositionality gap in language models. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 5687–5711. find
  23. Zhihong Shao, Yeyun Gong, Yelong Shen, Minlie Huang, Nan Duan, and Weizhu Chen. 2023. Enhancing retrieval-augmented large language models with iterative retrieval-generation synergy. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 9248–9274, Singapore. Association for…
  24. Weihang Su, Yichen Tang, Qingyao Ai, Zhijing Wu, and Yiqun Liu. 2024. DRAGIN: Dynamic retrieval augmented generation based on the real-time information needs of large language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers),…
  25. Harsh Trivedi, Niranjan Balasubramanian, Tushar Khot, and Ashish Sabharwal. 2022. musique: Multi-hop questions via single-hop question composition. Transactions of the Association for Computational Linguistics, 10:539–554. find
  26. Harsh Trivedi, Niranjan Balasubramanian, Tushar Khot, and Ashish Sabharwal. 2023. Interleaving retrieval with chain-of-thought reasoning for knowledge-intensive multi-step questions. In Proceedings of the 61st annual meeting of the association for computational linguistics (volume 1: long papers), p…
  27. Jia Zhu, Hanghui Guo, Weijie Shi, Zhangze Chen, and Pasquale De Meo. 2025a. Radio: Real-time hallucination detection with contextual index optimized query formulation for dynamic retrieval augmented generation. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 26129–… find
  28. Nana Zhu, Caihai Zhu, Qingfu Zhu, and Yuanxing Liu. 2025b. Style rewriter: A three-stage approach for stylistic generation of open-domain conversational responses. DATA INTELLIGENCE, 7(2):397–415. find
  29. 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.
  30. Fangyuan Xu, Weijia Shi, and Eunsol Choi. 2024. Recomp: Improving retrieval-augmented lms with context compression and selective augmentation. In The Twelfth International Conference on Learning Representations. find
  31. 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–…
  32. Zijun Yao, Weijian Qi, Liangming Pan, Shulin Cao, Linmei Hu, Liu Weichuan, Lei Hou, and Juanzi Li. 2025. SeaKR: Self-aware knowledge retrieval for adaptive retrieval augmented generation. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Paper…
  33. Biao Zhang, Yunwei Chen, and Daobin Gao. 2025. Development characteristics and topic analysis of international research on chatgpt. DATA INTELLI-GENCE, 7(4):1192–1217. find
  34. Xuhui Zheng, Kang An, Ziliang Wang, Yuhang Wang, and Yichao Wu. 2025. StepSearch: Igniting LLMs search ability via step-wise proximal policy optimization. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 21805–21830, Suzhou, China. Association for Comp…
  35. Wanjun Zhong, Lianghong Guo, Qiqi Gao, He Ye, and Yanlin Wang. 2024. Memorybank: Enhancing large language models with long-term memory. Proceedings of the AAAI Conference on Artificial Intelligence, 38(17):19724–19731. find
  36. This appendix provides supplementary material to ensure the clarity, depth, and reproducibility of our work. It is structured as follows: find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/shen-2026-acr-adaptive-context-refactoring-context/