Elamur Research⌕
☰
home / papers / 2026 / LoRE: Enhancing Search Relevance with Progressive Chain-of-Thought and Preference Alignment

LoRE: Enhancing Search Relevance with Progressive Chain-of-Thought and Preference Alignment

Paper · 2026 · Chenji Lu, Zhuo Chen, Hui Zhao et al.

Summary

The paper tackles e-commerce search relevance: deciding whether a retrieved item truly fits a user's query. The authors argue that existing LLM-based approaches lack a clear account of what such judgments require. They split the problem into three abilities: domain knowledge with reasoning, multimodal understanding, and adherence to business rules. Their framework, LoRE, trains a model in two stages.

  1. SFT "cold start". Chain-of-thought data is built step by step:
    • RAG-based query interpretation draws on titles of frequently clicked items and on merchant "selling points".
    • Relevance-focused image captions are added.
    • Final reasoning traces from a teacher LLM are grounded in explicit relevance rules.
  2. RL with verifiable rewards. This stage acts as a regularizer that trims unnecessary reasoning. It relies on GRPO, curriculum learning, dynamic sampling and clip-higher to prevent entropy collapse.

LoRE beats GPT-5 by 29.1% in Macro-F1 on the RAIR benchmark. Deployed on Taobao, it raised GoodRate by a cumulative 27% in relative terms. The work offers a practical template for domain-specific LLM post-training.

Key points

  • Labels follow RAIR's four-level scheme (L1–L2 irrelevant, L3–L4 relevant), and user intent is broken into 17 attribute dimensions. For query knowledge, Qwen3-235B filters high-CTR item titles and keeps 5 of them.
  • In the ablation on the General subset, adding rules gave the largest SFT gain (+9.0% Macro-F1). The final LoRE model reaches Acc@2 0.933, Acc@4 0.897 and Macro-F1 0.724.
  • CoT distillation raises the ceiling but hurts first-try accuracy. On the Hard subset, Pass@8 rises from 0.678 to 0.733 while Pass@1 falls from 0.671 to 0.648. RL then converts that headroom into reliable single-pass predictions.
  • Among the entropy-control options tested (on-policy updates, entropy regularization, clip-higher), clip-higher gave the best balance. Token-level importance sampling was also compared with sequence-level importance sampling, and the sequence-level variant sped up entropy collapse.
  • The online gains came from three sources:
    • caching LoRE scores: +4.8%
    • distilling LoRE into the production model with 60M samples: +9.5%
    • system and ranking strategy updates: +12.7%

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

🏢Alibaba GroupHangzhou · China168 works · 15,386 citations

 Fields, topics and concepts

📚 Shelves 5The 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 worksInformation Systems · 1,437 works

🧭 Topics 1Research subjects and languages the work is about.Reasoning · 1,426 works

Cite

Cite as Chenji Lu, Zhuo Chen, Hui Zhao, Zhiyuan Zeng, Gang Zhao, Junjie Ren, Lihaoran, Songyan Liu, Pengjie Wang, Chuan Yu, Jian Xu and Bo Zheng. 2026. LoRE: Enhancing Search Relevance with Progressive Chain-of-Thought and Preference Alignment. 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.1536

@inproceedings{lu2026lore,
  title = {{LoRE: Enhancing Search Relevance with Progressive Chain-of-Thought and Preference Alignment}},
  author = {Chenji Lu, Zhuo Chen, Hui Zhao, Zhiyuan Zeng, Gang Zhao, Junjie Ren, Lihaoran, Songyan Liu, Pengjie Wang, Chuan Yu, Jian Xu and Bo Zheng},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1536},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1536}
}

References

43 references, 15 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, Humen Zhong, Yuanzhi Zhu, Mingkun Yang, Zhaohai Li, Jianqiang Wan, Pengfei Wang, Wei Ding, Zheren Fu, Yiheng Xu, and 8 others. 2025. Qwen2.5-vl technical report. Preprint, arXiv:2502.13…
  2. Yoshua Bengio, Jérôme Louradour, Ronan Collobert, and Jason Weston. 2009. Curriculum learning. In ICML.
  3. David Carmel, Elad Haramaty, Arnon Lazerson, Liane Lewin-Eytan, and Yoelle Maarek. 2020. Why do people buy seemingly irrelevant items in voice product search? on the relation between product relevance and customer satisfaction in ecommerce. In Proceedings of the 13th International Conference on Web… find
  4. Zeyuan Chen, Haiyan Wu, Kaixin Wu, Wei Chen, Mingjie Zhong, Jia Xu, Zhongyi Liu, and Wei Zhang. 2025. Towards boosting LLMs-driven relevance modeling with progressive retrieved behavior-augmented prompting. In Proceedings of the 31st International Conference on Computational Linguistics: Industry Tr… find
  5. Gheorghe Comanici, Eric Bieber, Mike Schaekermann, Ice Pasupat, and Noveen Sachdeva. 2025. Gemini 2.5: Pushing the frontier with advanced reasoning, multimodality, long context, and next generation agentic capabilities. Preprint, arXiv:2507.06261.
  6. DeepSeek-AI, Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, and Ruoyu Zhang. 2025. Deepseekr1: Incentivizing reasoning capability in llms via reinforcement learning. Preprint, arXiv:2501.12948.
  7. Jacob Devlin. 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.
  8. Chenhe Dong, Shaowei Yao, Pengkun Jiao, Jianhui Yang, Yiming Jin, Zerui Huang, Xiaojiang Zhou, Dan Ou, and Haihong Tang. 2025. Taosr1: The thinking model for e-commerce relevance search. Preprint, arXiv:2508.12365. — [arXiv:2508.12365](https://arxiv.org/abs/2508.12365) find
  9. Zheng Fang, Donghao Xie, Ming Pang, Chunyuan Yuan, Xue Jiang, Changping Peng, Zhangang Lin, and Zheng Luo. 2025. Adore: Autonomous domain-oriented relevance engine for e-commerce. In Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGI… find
  10. Shangbin Feng, Weijia Shi, Yike Wang, Wenxuan Ding, Vidhisha Balachandran, and Yulia Tsvetkov. 2024. Don’t hallucinate, abstain: Identifying LLM knowledge gaps via multi-LLM collaboration. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Pape… find
  11. Yunye Gong, Robik Shrestha, Jared Claypoole, Michael Cogswell, Arijit Ray, Christopher Kanan, and Ajay Divakaran. 2024. BloomVQA: Assessing hierarchical multi-modal comprehension. In Findings of the Association for Computational Linguistics: ACL 2024, pages 14905–14918, Bangkok, Thailand. Associatio… find
  12. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, and Ahmad Al-Dahle. 2024. The llama 3 herd of models. Preprint, arXiv:2407.21783.
  13. Jujie He, Jiacai Liu, Chris Yuhao Liu, Rui Yan, Chaojie Wang, Peng Cheng, Xiaoyu Zhang, Fuxiang Zhang, Jiacheng Xu, Wei Shen, Siyuan Li, Liang Zeng, Tianwen Wei, Cheng Cheng, Bo An, Yang Liu, and Yahui Zhou. 2025a. Skywork open reasoner 1 technical report. Preprint, arXiv:2505.22312. find
  14. Mingqian He, Fei Zhao, Chonggang Lu, Ziyan Liu, Yue Wang, and Haofu Qian. 2025b. Gencls++: Pushing the boundaries of generative classification in llms through comprehensive sft and rl studies across diverse datasets. Preprint, arXiv:2504.19898. — [arXiv:2504.19898](https://arxiv.org/abs/2504.19898) find
  15. Christian Herold, Michael Kozielski, Tala Bazazo, Pavel Petrushkov, Yannick Versley, Seyyed Hadi Hashemi, Patrycja Cieplicka, Dominika Basaj, and Shahram Khadivi. 2025. Domain adaptation of foundation LLMs for e-commerce. In Proceedings of the 63rd Annual Meeting of the Association for Computational… find
  16. Po-Sen Huang, Xiaodong He, Jianfeng Gao, and 1 others. 2013. Learning deep structured semantic models for web search using clickthrough data. In Proceedings of the 22nd ACM international conference on Information & Knowledge Management, pages 2333– 2338. find
  17. Samuel Humeau, Kurt Shuster, Marie-Anne Lachaux, and Jason Weston. 2019. Poly-encoders: Architectures and pre-training strategies for fast and accurate multi-sentence scoring. In International Conference on Learning Representations. find
  18. Lei Li, Yuqi Wang, Runxin Xu, Peiyi Wang, Xiachong Feng, Lingpeng Kong, and Qi Liu. 2024. Multimodal ArXiv: A dataset for improving scientific comprehension of large vision-language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Pap…
  19. Ming Li, Jike Zhong, Shitian Zhao, Yuxiang Lai, Haoquan Zhang, Wang Bill Zhu, and Kaipeng Zhang. 2025a. Think or not think: A study of explicit thinking in rule-based visual reinforcement fine-tuning. Preprint, arXiv:2503.16188. — [arXiv:2503.16188](https://arxiv.org/abs/2503.16188) find
  20. Moxin Li, Yong Zhao, Wenxuan Zhang, Shuaiyi Li, Wenya Xie, See-Kiong Ng, Tat-Seng Chua, and Yang Deng. 2025b. Knowledge boundary of large language models: A survey. Preprint, arXiv:2412.12472.
  21. Hong Liu, Saisai Gong, Yixin Ji, Kaixin Wu, Jia Xu, and Jinjie Gu. 2024. Boosting llm-based relevance modeling with distribution-aware robust learning. In Proceedings of the 33rd ACM International Conference on Information and Knowledge Management, CIKM ’24, page 4718–4725, New York, NY, USA. Associ… find
  22. Huanghai Liu, Quzhe Huang, Qingjing Chen, Yiran Hu, Jiayu Ma, Yun Liu, Weixing Shen, and Yansong Feng. 2025. JUREX-4E: Juridical expert-annotated four-element knowledge base for legal reasoning. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 3794– 38… find
  23. Yinhan Liu. 2019. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692.
  24. Ziyang Liu, Chaokun Wang, Hao Feng, Lingfei Wu, and Liqun Yang. 2022. Knowledge distillation based contextual relevance matching for e-commerce product search. Preprint, arXiv:2210.01701. find
  25. Chenji Lu, Zhuo Chen, Hui Zhao, Zhenyi Wang, Pengjie Wang, Jian Xu, and Bo Zheng. 2025. Rair: A rule-aware benchmark uniting challenging long-tail and visual salience subset for e-commerce relevance assessment. Preprint, arXiv:2512.24943. — [arXiv:2512.24943](https://arxiv.org/abs/2512.24943) find
  26. Navid Mehrdad, Hrushikesh Mohapatra, Mossaab Bagdouri, Prijith Chandran, Alessandro Magnani, Xun-fan Cai, Ajit Puthenputhussery, Sachin Yadav, Tony Lee, ChengXiang Zhai, and Ciya Liao. 2024. Large language models for relevance judgment in product search. Preprint, arXiv:2406.00247. — [arXiv:2406.00247](https://arxiv.org/abs/2406.00247) find
  27. Qwen, :, An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, Huan Lin, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jingren Zhou, and 25 others. 2025. Qwen2.5 technical report. Preprint, arXiv:2412.15115. find
  28. Rafael Rafailov, Archit Sharma, Eric Mitchell, Stefano Ermon, Christopher D. Manning, and Chelsea Finn. 2024. Direct preference optimization: Your language model is secretly a reward model. Preprint, arXiv:2305.18290.
  29. N Reimers. 2019. Sentence-bert: Sentence embeddings using siamese bert-networks. arXiv preprint arXiv:1908.10084.
  30. Stephen E. Robertson and Steve Walker. 1994. Some simple effective approximations to the 2-poisson model for probabilistic weighted retrieval. In Annual International ACM SIGIR Conference on Research and Development in Information Retrieval. find
  31. Karen Sparck Jones. 1988. A statistical interpretation of term specificity and its application in retrieval, page 132–142. Taylor Graham Publishing, GBR. find
  32. Tian Tang, Zhixing Tian, Zhenyu Zhu, Chenyang Wang, Haiqing Hu, Guoyu Tang, Lin Liu, and Sulong Xu. 2025. Lref: A novel llm-based relevance framework for e-commerce search. In Companion Proceedings of the ACM on Web Conference 2025, WWW ’25, page 468–475, New York, NY, USA. Association for Computing… find
  33. Kimi Team, Angang Du, Bofei Gao, Bowei Xing, Changjiu Jiang, Cheng Chen, Cheng Li, and Chenjun Xiao. 2025. Kimi k1.5: Scaling reinforcement learning with llms. Preprint, arXiv:2501.12599.
  34. Gang Zhao, Ximing Zhang, Chenji Lu, Hui Zhao, Tianshu Wu, Pengjie Wang, Jian Xu, and Bo Zheng. 2025a. Explainable llm-driven multi-dimensional distillation for e-commerce relevance learning. In Companion Proceedings of the ACM on Web Conference 2025, WWW ’25, page 631–640, New York, NY, USA. Associa… find
  35. Yuzhong Zhao, Yue Liu, Junpeng Liu, Jingye Chen, Xun Wu, Yaru Hao, Tengchao Lv, Shaohan Huang, Lei Cui, Qixiang Ye, Fang Wan, and Furu Wei. 2025b. Geometric-mean policy optimization. Preprint, arXiv:2507.20673. — [arXiv:2507.20673](https://arxiv.org/abs/2507.20673) find
  36. Chujie Zheng, Shixuan Liu, Mingze Li, Xiong-Hui Chen, Bowen Yu, Chang Gao, Kai Dang, Yuqiong Liu, Rui Men, An Yang, Jingren Zhou, and Junyang Lin. 2025. Group sequence policy optimization. Preprint, arXiv:2507.18071.
  37. Qihuang Zhong, Liang Ding, Xiantao Cai, Juhua Liu, Bo Du, and Dacheng Tao. 2025. KaFT: Knowledge-aware fine-tuning for boosting LLMs’ domain-specific question-answering performance. In Findings of the Association for Computational Linguistics: ACL 2025, pages 24085–24100, Vienna, Austria. Associatio… find
  38. Wei Wang, Bin Bi, Ming Yan, Chen Wu, Zuyi Bao, Jiangnan Xia, Liwei Peng, and Luo Si. 2019. Structbert: Incorporating language structures into pre-training for deep language understanding. arXiv preprint arXiv:1908.04577. find
  39. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, and Bo Zheng. 2025a. Qwen3 technical report. Preprint, arXiv:2505.09388.
  40. Haoqiang Yang, Congde Yuan, Kun Bai, Mengzhuo Guo, Wei Yang, and Chao Zhou. 2025b. Hit model: A hierarchical interaction-enhanced two-tower model for pre-ranking systems. Preprint, arXiv:2505.19849. — [arXiv:2505.19849](https://arxiv.org/abs/2505.19849) find
  41. Qiying Yu, Zheng Zhang, Ruofei Zhu, Yufeng Yuan, Xiaochen Zuo, Yu Yue, Weinan Dai, Tiantian Fan, Gaohong Liu, Lingjun Liu, Xin Liu, Haibin Lin, Zhiqi Lin, Bole Ma, Guangming Sheng, Yuxuan Tong, Chi Zhang, Mofan Zhang, Wang Zhang, and 16 others. 2025. Dapo: An open-source llm reinforcement learning s…
  42. Chunyuan Yuan, Yiming Qiu, Mingming Li, Haiqing Hu, Songlin Wang, and Sulong Xu. 2023. A multi-granularity matching attention network for query intent classification in e-commerce retrieval. In Companion Proceedings of the ACM Web Conference 2023, WWW ’23, page 416–420. ACM. find
  43. Wei Zhang, Miaoxin Cai, Tong Zhang, Yin Zhuang, and Xuerui Mao. 2024. Earthgpt: A universal multimodal large language model for multisensor image comprehension in remote sensing domain. IEEE Transactions on Geoscience and Remote Sensing, 62:1–20. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/lu-2026-lore-enhancing-search-relevance-progressive/