publications

2026

  1. Preprint
    ALIBI: Reasoning Failure Attacks on LLM-Based Vulnerability Detectors via Adversarial Code Comments
    Zixuan Wu, and Cristina Nita-Rotaru
    Preprint. In submission to USENIX 2027, 2026
    arXiv
  2. LMPL
    The Best Programming Language for Tokenmaxxing: An Investigation of Coding Agent Behavior Across Programming Languages
    Zixuan Wu, Carolyn Jane Anderson, and Arjun Guha
    In International Workshop on Language Models and Programming Languages (LMPL) at SPLASH 2026 , 2026
    arXiv
  3. EMNLP Findings
    AgentPack: A Dataset of Code Changes, Co-Authored by Agents and Humans
    Yangtian Zi, Zixuan Wu, Aleksander Boruch-Gruszecki, and 2 more authors
    In Findings of the 2026 Conference on Empirical Methods in Natural Language Processing , 2026
    arXiv
  4. ICLR
    Agnostics: Learning to Synthesize Code in Any Programming Language with a Universal Reinforcement Learning Environment
    Aleksander Boruch-Gruszecki, Yangtian Zi, Zixuan Wu, and 4 more authors
    In The Fourteenth International Conference on Learning Representations , 2026
    arXiv

2025

  1. AACL-IJCNLP
    ReasoningWeekly: A General Knowledge and Verbal Reasoning Challenge for Large Language Models
    Zixuan Wu, Francesca Lucchetti, Aleksander Boruch-Gruszecki, and 5 more authors
    In The 14th International Joint Conference on Natural Language Processing and the 4th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Oral) , 2025
    arXiv
  2. NAACL
    substancebeatsstyle0.png
    Substance Beats Style: Why Beginning Students Fail to Code with LLMs
    Francesca Lucchetti, Zixuan Wu, Arjun Guha, and 2 more authors
    In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics (Oral) , 2025
    arXiv HTML
  3. ACL Findings
    glyph0.png
    GlyphPattern: An Abstract Pattern Recognition for Vision-Language Models
    Zixuan Wu, Yoolim Kim, and Carolyn Jane Anderson
    In Findings of the Association for Computational Linguistics: ACL 2025 , 2025
    arXiv HTML