Xiaosong Yuan苑小松

News

• Feb 2026 New Joined Tongyi Lab, Alibaba as a research intern, advised by Yuexiang Xie and Yaliang Li.
• Jan 2026 Three papers accepted to ICLR 2026 (one oral, one poster).
• Jan 2026 SeaRAG accepted to WWW 2026.
• Sep 2025 Two papers accepted to EMNLP 2025 (oral).
• Sep 2024 Instance-adaptive zero-shot chain-of-thought prompting accepted to NeurIPS 2024.
• Dec 2023 Began research internship at Apsara Lab (former City Brain Lab, DAMO Academy, Tongyi Lab-BigData & Intelligence Analysis Lab), Alibaba, advised by Chen Shen and Prof. Jieping Ye.

Publications

First-author papers are highlighted. Full list also on Google Scholar.

• WWW 2026

  SeaRAG: Reducing Hallucination in Retrieval-Augmented Generation via Statement-Entity Adaptive Ranking

  X Yuan, X Zhang, D Zhao, Y Zhang, Y Wang.

  The ACM Web Conference (WWW), pp. 2339–2348, 2026.

  ACM DL

• ICLR 2026

  Differential Fine-Tuning Large Language Models Towards Better Diverse Reasoning Abilities

  X Yuan, C Shen, S Yan, K Liu, X Zhang, L Sinan Fan, Xie, W Wang, et al.

  International Conference on Learning Representations (ICLR), 2026.

  OpenReview

• ICLR · Oral 2026

  Hallucination Begins Where Saliency Drops

  X Zhang, Y Zhu, C Gu, X Yuan, Q Zhao, J Cao, F Tang, S Fan, Y Shen, et al.

  International Conference on Learning Representations (ICLR), 2026.

  OpenReview

• ICLR 2026

  Context Tokens are Anchors: Understanding the Repetition Curse in dMLLMs from an Information Flow Perspective

  Q Zhao, X Zhang, S Chang, Q Chen, X Yuan, X Chen, L Liu, J Zhang, et al.

  International Conference on Learning Representations (ICLR), 2026.

  OpenReview

• arXiv 2026

  Reasoning Fails Where Step Flow Breaks

  X Xu, Y Pan, X Yuan, Z Shen, M Su, Y Su, X Zhang.

  arXiv preprint arXiv:2604.06695, 2026.

  arXiv

• arXiv 2026

  ART: Attention Replacement Technique to Improve Factuality in LLMs

  Z Luo, Y Quan, X Zhang, X Yuan, C Shen.

  arXiv preprint arXiv:2604.06393, 2026.

  arXiv

• arXiv 2026

  On the Step Length Confounding in LLM Reasoning Data Selection

  B Wang, R Miao, C Shen, S Yan, K Liu, X Li, X Yuan, S Fan, J Zhang, J Ye.

  arXiv preprint arXiv:2604.06834, 2026.

  arXiv

• IEEE TCE 2025

  LSFusion: Ladder-Side Attribute Composition for Multi-Aspect Controllable Text Generation

  X Yuan, et al.

  IEEE Transactions on Consumer Electronics, 2025.

  IEEE Xplore

• EMNLP · Oral 2025

  Shallow Focus, Deep Fixes: Enhancing Shallow Layers Vision Attention Sinks to Alleviate Hallucination in LVLMs

  X Zhang, Y Quan, C Shen, C Gu, X Yuan, S Yan, J Cao, H Cheng, K Wu, et al.

  Conference on Empirical Methods in Natural Language Processing (EMNLP), 2025.

  ACL Anthology

• EMNLP · Oral 2025

  Seeing clearly by layer two: Enhancing attention heads to alleviate hallucination in lvlms

  X Zhang, Y Quan, C Gu, C Shen, X Yuan, S Yan, H Cheng, K Wu, J Ye.

  Conference on Empirical Methods in Natural Language Processing (EMNLP), 2025. arXiv:2411.09968.

  arXiv

• NAACL · Oral 2025

  From redundancy to relevance: Enhancing explainability in multimodal large language models

  X Zhang, C Shen, X Yuan, S Yan, L Xie, W Wang, C Gu, H Tang, J Ye.

  North American Chapter of the ACL (NAACL), 2025.

  ACL Anthology

• ACM MM 2025

  Mca-llava: Manhattan causal attention for reducing hallucination in large vision-language models

  Q Zhao, X Zhang, Y Li, Y Xing, X Yuan, F Tang, S Fan, X Chen, DH Wang, et al.

  ACM International Conference on Multimedia (ACM MM), pp. 3981–3990, 2025.

  ACM DL

• ICLR 2025

  Improving Complex Reasoning with Dynamic Prompt Corruption: A soft prompt Optimization Approach

  S Fan, L Xie, C Shen, G Teng, X Yuan, X Zhang, C Huang, W Wang, X He, et al.

  International Conference on Learning Representations (ICLR), 2025.

  OpenReview

• ESWA 2025

  DRIVE: An adjustable parallel architecture based on evidence awareness for fake news detection

  M Cong, L Liu, J Han, B Zhu, X Yuan, T Peng.

  Expert Systems with Applications, vol. 266, 126043, 2025.

  ScienceDirect

• arXiv 2025

  Efficient reasoning through suppression of self-affirmation reflections in large reasoning models

  K Liu, C Shen, Z Zhang, J Liu, X Yuan.

  arXiv preprint arXiv:2506.12353, 2025.

  arXiv

• NeurIPS 2024

  Instance-adaptive zero-shot chain-of-thought prompting

  X Yuan, C Shen, S Yan, X Zhang, L Xie, W Wang, R Guan, Y Wang, J Ye.

  Advances in Neural Information Processing Systems (NeurIPS), vol. 37, pp. 125469–125486, 2024.

  NeurIPS

• PAKDD · Oral 2023

  The causal strength bank: a new benchmark for causal strength classification

  X Yuan, R Guan, W Zuo, Y Zhang.

  Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD), pp. 111–122, 2023.

  Springer

• IJCNN · Oral 2023

  TC-GAT: Graph Attention Network for Temporal Causality Discovery

  X Yuan, K Chen, W Zuo, Y Zhang.

  International Joint Conference on Neural Networks (IJCNN), 2023.

  IEEE Xplore

Experience

• Feb 2026 – Present
  Research Intern, Tongyi Lab, Alibaba Group Advised by Yuexiang Xie and Yaliang Li. Working on LLM agent systems.
• Jan 2026 – Feb 2026
  Research Intern, AQ Team, Ant Group
• Dec 2023 – Dec 2025
  Research Intern, City Brain Lab, DAMO Academy, Alibaba Advised by Chen Shen and Prof. Jieping Ye. Topics: Improving LLM reasoning.

Selected Industrial Projects

• QwenPaw — Tongyi Lab, Alibaba Cloud (Feb 2026 – Present).
• X1 Project — Apsara Lab, Alibaba Cloud (Oct 2024 – 2025).

Education

• Sep 2021 – Present
  Ph.D. Candidate, College of Computer Science and Technology, Jilin University Successive master–doctor program. Advisor: Prof. Renchu Guan.
• Sep 2019 – Jun 2021
  M.Sc. Candidate, College of Computer Science and Technology, Jilin University Successive master–doctor program. Research under Prof. Wanli Zuo.
• Sep 2014 – Jun 2018
  B.Eng., College of Software, Jilin University

Academic Service

Honors & Awards

• Academic Performance Scholarship, Jilin University — 2021, 2022, 2023.

Skills