publications

2026

  1. Preprint
    VisAudit: Evaluating Multimodal Agents for Visual Diagnosis and Repair
    Shicheng Liu, Adam Kahirov, Qi Zhang, and 5 more authors
    In submission to ICLR 2027, 2026
    HTML
  2. NeurIPS 2026
    Position: Mechanistic Interpretability Needs a Logic of Inference
    Zhimin Hu, Chen Jiang, Hanbo Xie, and 2 more authors
    NeurIPS - Interp4Discovery (🌟Oral), 2026
    HTML
  3. ICML 2026
    Language Models Represent and Transform Concepts with Shared Geometry
    Zhimin Hu, Lanhao Niu, and Sashank Varma
    ICML - Mech Interp (🌟Spotlight), 2026
    HTML
  4. CogSci 2026
    The Representational Geometry of Number
    Zhimin Hu, Lanhao Niu, and Sashank Varma
    In Proceedings of the Annual Meeting of the Cognitive Science Society 48, 2026
    HTML
  5. CogSci 2026
    Are More Tokens Rational? Inference-Time Scaling in Language Models as Adaptive Resource Rationality
    Zhimin Hu, Riya Roshan, and Sashank Varma
    In Proceedings of the Annual Meeting of the Cognitive Science Society 48, 2026
    HTML

2025

  1. ICLR 2025
    LLMs Are Not Good Strategists, Yet Memory-Enhanced Agency Boosts Reasoning
    Zhimin Hu* and Yi Wu*
    In ICLR - Reasoning and Planning for LLM, 2025
    * equal contribution
    HTML
  2. Neural Comp
    From Function to Implementation: Exploring Degeneracy in Evolved Artificial Agents
    Zhimin Hu, Oğulcan Cingiler, Clifford Bohm, and 1 more author
    Neural Computation (🌟Featured on Cover), 2025
    HTML

2024

  1. Evolang XV
    Failures and Successes to Learn a Core Conceptual Distinction From the Statistics of Language
    Zhimin Hu, Jeroen Van Paridon, and Gary Lupyan
    In The Evolution of Language: Proceedings of the 15th International Conference, 2024
    HTML