Rongwu Xu (许融武)

About [中文]

Education

• Master in Computer Science, 2022-2025
  Institute for Interdisciplinary Information Sciences, Tsinghua University
  Advisor: Prof. Wei Xu
• Bachelor in Computer Science, 2018-2022
  Department of Computer Science and Technology, Tsinghua University
• Visiting Student Researcher, Summer 2021
  Department of Computer Science, Duke University

Research

My research interests lie in Natural Language Processing (NLP) and Large Language Models (LLMs). In general, I am always fascinated by the notion of thinking of AI models as a form of approximation to human cognition and social interaction beyond merely functional analogs. My previous work has focused on comprehending the interaction of language models with the context where human users engage. I have also worked on controlled NLG and evaluation frameworks that emulate human assessment techniques. My ultimate goal is to design language intelligence capable of thinking and acting like humans and augmenting human capabilities.

The following research topics, therefore, pique my interest:

1. Human-AI Alignment: What are the key ingredients to make AI systems be effectively aligned with human values, behavioral patterns and expectations beyond machine learning algorithms?
2. Machine Psychology: What are the key similarities and differences between AI models and human mind? How to use psychology-inspired behavioral experiments to test and understand LLMs?

1. Science and Practice of Evaluation: What are the considerations for establishing a rigorous science and practice of evaluation that goes beyond task-driven approaches?
2. LLM as Tools and Domain Applications: In what ways can LLMs be leveraged as tools across various domains, especially specialized ones? E.g., computational social science.
3. AI Ethics and Safety: What are the primary ethical and safety concerns associated with the deployment and application of LLMs, and how can these be addressed to ensure responsible and trustworthy use?

News

Research in the rapidly evolving field of AI/NLP can be challenging for newcomers. If you are interested in my research or have preliminary ideas you'd like to explore, I'm more than willing to offering guidance. We can work towards submitting papers to top venues such as ACL/EMNLP/NAACL. Feel free to drop me an Email if interested.

Publications

See Google Scholar for a complete list and bibliometric.

Preprints

1. DebateQA: Evaluating Question Answering on Debatable Knowledge
   Rongwu Xu*, Xuan Qi*, Zehan Qi, Wei Xu, Zhijiang Guo
   arXiv Preprint (major revision at TACL)
   [Code]
   
2. On the Role of Attention Heads in Large Language Model Safety
   Zhenhong Zhou, Haiyang Yu, Xinghua Zhang, Rongwu Xu, Fei Huang, Kun Wang, Yang Liu, Junfeng Fang, Yongbin Li
arXiv Preprint

Conference Proceedings

1. Course-Correction: Safety Alignment Using Synthetic Preferences
   Rongwu Xu*, Yishuo Cai*, Zhenhong Zhou, Renjie Gu, Haiqin Wang, Yan Liu, Tianwei Zhang, Wei Xu, Han Qiu
   EMNLP 2024 (Industry)
   [Code][Poster]
   
2. MR-BEN: A Comprehensive Meta-Reasoning Benchmark for Large Language Models
   Zhongshen Zeng, Yinhong Liu, Yingjia Wan, Jingyao Li, Pengguang Chen, Jianbo Dai, Yuxuan Yao, Rongwu Xu, Zehan Qi, Wanru Zhao, Linling Shen, Jianqiao Lu, Haochen Tan, Yukang Chen, Hao Zhang, Zhan Shi, Bailin Wang, Zhijiang Guo, Jiaya Jia
   NeurIPS 2024
   [Code][Project Page]
   
3. Knowledge Conflicts for LLMs: A Survey
   Rongwu Xu*, Zehan Qi*, Zhijiang Guo, Cunxiang Wang, Hongru Wang, Yue Zhang, Wei Xu
   EMNLP 2024 (Main)
   [Code][机器之心][Talk (Chinese)][Slides][Poster]
   
4. $LONG^{2}RAG$: Evaluating Long-Context & Long-Form Retrieval-Augmented Generation with Key Point Recall
   Zehan Qi*, Rongwu Xu*, Zhijiang Guo, Cunxiang Wang, Hao Zhang, Wei Xu
   EMNLP 2024 (Findings)
   
5. Walking in Others' Shoes: How Perspective-Taking Guides Large Language Models in Reducing Toxicity and Bias
   Rongwu Xu, Zi'an Zhou, Tianwei Zhang, Zehan Qi, Su Yao, Ke Xu, Wei Xu, Han Qiu
   EMNLP 2024 (Main)
   [Poster]
   
6. Sing it, Narrate it: Quality Musical Lyrics Translation
   Zhuorui Ye, Jinhan Li, Rongwu Xu
   EMNLP 2024 (Findings)
   
7. How Alignment and Jailbreak Work: Explain LLM Safety through Intermediate Hidden States
   Zhenhong Zhou, Haiyang Yu, Xinghua Zhang, Rongwu Xu, Fei Huang, Yongbin Li
   EMNLP 2024 (Findings)
   [Code][Poster]
   
8. Preemptive Answer ``Attacks'' on Chain-of-Thought Reasoning
   Rongwu Xu*, Zehan Qi*, Wei Xu
   ACL 2024 (Findings)
   [Code][Poster]
   
9. The Earth is Flat because...: Investigating LLMs' Belief towards Misinformation via Persuasive Conversation
   Rongwu Xu, Brian S. Lin, Shujian Yang, Tianqi Zhang, Weiyan Shi, Tianwei Zhang, Zhixuan Fang, Wei Xu, Han Qiu
   ACL 2024 (Oral, Main)
   🏆 Outstanding Paper Award [Certificate]
   [Code][机器之心][Project Page][Video][Poster]
   
10. Exploring Chinese Humor Generation: A Study on Two-Part Allegorical Sayings
    Rongwu Xu
    IJCNN 2024
    
11. Tempo: Confidentiality Preservation in Cloud-Based Neural Network Training
    Rongwu Xu and Zhixuan Fang
    IJCNN 2024
    
12. MISO: Legacy-compatible Privacy-preserving Single Sign-on using Trusted Execution Environments
    Rongwu Xu, Sen Yang, Fan Zhang, Zhixuan Fang
    EuroS&P 2023
    [Project Page]
    
13. LSync: A Universal Event-synchronizing Solution for Live Streaming
    Yifan Xu, Fan Dang, Rongwu Xu, Xinlei Chen, Yunhao Liu
    INFOCOM 2022
    
14. LifeRec: A Mobile App for Lifelog Recording and Ubiquitous Recommendation
    Jiayu Li, Hantian Zhang*, Zhiyu He*, Rongwu Xu*, Pingfei Wu*, Min Zhang, Yiqun Liu, Shaoping Ma
    CHIIR 2022
    [Code]
    

Journal Articles

1. LSync: A Universal Timeline-synchronizing Solution for Live Streaming
   Fan Dang*, Yifan Xu*, Rongwu Xu, Xinlei Chen, Yunhao Liu
   IEEE/ACM Trans. on Networking, 2024
   

* Equal Contribution, ^ Corresponding Author

Awards*

• Most Recognized Research Outcomes at Tsinghua University Nomination (清华大学最受师生关注的年度亮点成果提名, Top-2 in IIIS), 2024
• Tsinghua University Excellent Teaching Assistant (清华大学优秀助教, Top 2%), 2024
• National Scholarship (国家奖学金, Top 1%), 2024
• Tsinghua University Outstanding Student Cadre (清华大学优秀学生干部, Top 1.5%, 121 out of 9000+), 2024
• ACL 2024 Outstanding Paper Award (Top 0.79%, 35 out of 4407), 2024
• Tsinghua-Yangtze River Delta International R&D Community Talent Scholarship, 2023
• Tsinghua University Overall Excellence Scholarship (Top 10%), 2023
• Tsinghua University Overall Excellence Scholarship (Top 10%), 2022
• Tsinghua University Technological Innovation Excellence Scholarship, 2020
• Tsinghua-Panasonic Scholarship (Top 10%), 2019
• Outstanding Volunteers in Beijing, 2018

* Only honors recognized at or above the university level are shown.

Experience

• Research Intern, Sep 2024-Present
  Topic: Clinical Large Multimodal (Language) Models
  Shanghai Qi Zhi Institute, Shanghai, China
• Research Intern, Apr 2024-Aug 2024
  Topic: Generative Multimodal Models
  TongYi Vision Intelligence Lab, Alibaba Inc., Beijing, China
• Research Intern, Dec 2022-Jan 2023
  Topic: Decentralized Finance
  Shanghai Qi Zhi Institute, Shanghai, China

Activity

Professional

• Peer Reviewer: ACL Rolling Review (2024-Present)

Talks

• “做科研的选择” (在2024年交叉信息院开学典礼上作为在校生代表的发言), IIIS, Tsinghua University, Aug 2024
  [Speech (in Chinese)]
• Knowledge Conflicts for LLMs, NICE (w. Soochow University), Jul 2024
  [2-Hour Talk (in Chinese)][Slides]
  
• LLMs Safety Vulnerabilities via Contextual Misinformation, IIIS, Tsinghua University, May 2024
  [3-min Clip]

Teaching

I was invited to share my teaching experience (see slides in Chinese) within my department (2024 & 2023) and awarded as excellent teaching assistant at Tsinghua University (2024).

• Head Teaching Assistant and Organizer, Introduction to Large Language Model Applications, IIIS, Tsinghua University, Spring 2025 & Spring 2024
  Check out the code we developed for the course. (10+contributers, involing LLM-based App design, inference, efficient fine-tuning, etc.)
• Teaching Assistant, Operating Systems and Distributed Systems, IIIS, Tsinghua University, Fall 2023
  
• Teaching Assistant, Distributed Systems, IIIS, Tsinghua University, Spring 2022
  

Miscellaneous