Lechen Zhang

prof_pic.jpg

Hi, I am Lechen! I am a first-year Ph.D. Student in Computer Science at University of Illinois Urbana-Champaign, advised by Professor Tal August. I have broad interests in Natural Language Processing and Machine Learning. I am particularly interested in the generalization and robustness of large language models (LLMs), especially their efficient learning and reasoning capabilities in out-of-domain settings. My recent research focuses on evaluating and improving LLM personalization, with the goal of aligning general model behavior more effectively with downstream tasks and user needs.

Previously, I earned my master’s degree from University of Michigan and my bachelor’s degree from Shanghai Jiao Tong University. During my time at Michigan, I was a member of Blablablab advised by Professor David Jurgens, and also worked with Professor Lu Wang as part of the Launch Lab.

You can find my CV here.

news

Nov 04, 2025 I will attend EMNLP 2025 and present our paper VeriFact! Can’t wait to visit Suzhou and meet friends!
Apr 15, 2025 I’m excited to share that I’ll be joining UIUC this fall. Looking forward to meeting new friends!
Nov 12, 2024 Excited to Attend EMNLP 2024! See you in Miami :palm_tree:
Oct 15, 2024 Submitting 4 new papers in this cycle! Check out our newly released preprints about System Prompting, Factuality Evaluation, and Human-AI Dialog Simulation.
Jun 18, 2024 I am presenting our paper at NAACL 2024 today. Check our slides here!

selected publications

* indicates equal contribution
  1. Preprint
    Under Review
    multilingual_sys_prompt_preview.jpg
    Cross-Lingual Prompt Steerability: Towards Accurate and Robust LLM Behavior across Languages
    Lechen Zhang*, Yusheng Zhou*, Tolga Ergen, Lajanugen Logeswaran, Moontae Lee, and David Jurgens
    arXiv preprint arXiv:2512.02841, Dec 2025
  2. MATH-AI @ NIPS 2025
    Under Review
    skill_data_select_preview.jpg
    Skill-Aware Data Selection and Fine-Tuning for Data-Efficient Reasoning Distillation
    Lechen Zhang, Yunxiang Zhang, Wei Hu, and Lu Wang
    In The 5th Workshop on Mathematical Reasoning and AI at NeurIPS 2025, Dec 2025
  3. VeriFact: Enhancing Long-Form Factuality Evaluation with Refined Fact Extraction and Reference Facts
    Xin Liu, Lechen Zhang, Sheza Munir, Yiyang Gu, and Lu Wang
    In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, Nov 2025
  4. SCALR @ COLM 2025
    Under Review
    thinklogit_preview.jpg
    Logit Arithmetic Elicits Long Reasoning Capabilities Without Training
    Yunxiang Zhang, Muhammad Khalifa, Lechen Zhang, Xin Liu, Ayoung Lee, Xinliang Frederick Zhang, Farima Fatahi Bayat, and Lu Wang
    In The 1st Workshop on Test-time Scaling and Reasoning Models at COLM 2025, Oct 2025
  5. ACL 2025
    Oral
    factbench_preview.jpg
    FactBench: A Dynamic Benchmark for In-the-Wild Language Model Factuality Evaluation
    Farima Fatahi Bayat, Lechen Zhang, Sheza Munir, and Lu Wang
    In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Jul 2025
  6. Toward Global AI Inclusivity: A Large-Scale Multilingual Terminology Dataset (GIST)
    Jiarui Liu, Iman Ouzzani, Wenkai Li, Lechen Zhang, Tianyue Ou, Houda Bouamor, Zhijing Jin, and Mona T. Diab
    In Findings of the Association for Computational Linguistics: ACL 2025, Jul 2025
  7. Causally Modeling the Linguistic and Social Factors that Predict Email Response
    Yinuo Xu*, Hong Chen*, Sushrita Rakshit*, Aparna Ananthasubramaniam*, Omkar Yadav*, Mingqian Zheng*, Michael Jiang*Lechen Zhang*, Bowen Yi*, Kenan Alkiek*, Abraham Israeli*, Bangzhao Shu*, Hua Shen*, Jiaxin Pei*, Haotian Zhang*, Miriam Schirmer*, and David Jurgens
    In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), Apr 2025
  8. Preprint
    Under Review
    sprig_preview.jpg
    SPRIG: Improving Large Language Model Performance by System Prompt Optimization
    Lechen Zhang, Tolga Ergen, Lajanugen Logeswaran, Moontae Lee, and David Jurgens
    arXiv preprint arXiv:2410.14826, Oct 2024
  9. Under Review
    latent_geo_preview.jpg
    Latent Geographies: Joint Embeddings of Text and Visual Cues for Social Media Geolocation
    Lechen Zhang*, Abraham Israeli*, Rohan Raju, and David Jurgens
    Oct 2024
  10. Preprint
    Under Review
    human_simulation_preview.jpg
    Real or Robotic? Assessing Whether LLMs Accurately Simulate Qualities of Human Responses in Dialogue
    Jonathan Ivey*, Shivani Kumar*, Jiayu Liu*, Hua Shen*, Sushrita Rakshit*, Rohan Raju*, Haotian Zhang*, Aparna Ananthasubramaniam*, Junghwan Kim*, Bowen Yi*, Dustin Wright*, Abraham Israeli*, Anders Giovanni Møller*Lechen Zhang*, and David Jurgens
    arXiv preprint arXiv:2409.08330, Sep 2024
  11. NAACL 2024
    Oral
    llm_personas_preview.jpg
    You don’t need a personality test to know these models are unreliable: Assessing the Reliability of Large Language Models on Psychometric Instruments
    Bangzhao Shu*Lechen Zhang*, Minje Choi, Lavinia Dunagan, Lajanugen Logeswaran, Moontae Lee, Dallas Card, and David Jurgens
    In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics (NAACL), Jun 2024