avatar

Huey Sun

huey3sn (at) gmail.com


About Me

I’m a researcher interested in AI agents, hierarchical reasoning, and preference-based systems. Previously, I was a Research Scholar at Cohere Labs, where I worked on personalization and multi-adapter architectures. Before that, I was at UCL, Microsoft, and Pomona College. I love jazz piano, my dog Coco, and noodle soups (especially 麵線).

Selected Work

  1. Disentangling Preference Soups With Parameter-Efficient Personalization
    Huey Sun, Shivalika Singh, Arash Ahmadian, Jon Ander Campos, Beyza Ermis, Sara Hooker
    Manuscript in preparation (2026).

  2. arXiv
    Enhancing Instruction-Following Capabilities in Seq2Seq Models: DoLA Adaptations for T5
    Huey Sun, Anabel Yong, Lorenzo Gilly, Felipe Jin
    arXiv preprint (2025).
    PDF Code   Preprint

  3. arXiv
    Diminishing Returns in Self-Supervised Learning
    Oliver Bridge, Huey Sun, Botond Branyicskai-Nagy, Charles-Antoine D’Ornano, Shomit Basu
    arXiv preprint (2025).
    PDF Code   Preprint

  4. Thesis
    Thinking Forwards, Backwards, and in Code: Improving Tool Acquisition in Large Language Models
    Huey Sun, Yihong Chen, Pontus Stenetorp
    MsC Thesis (2024).
    PDF Code

  5. LOD
    Limits of Transfer Learning
    Jake Williams, Abel Tadesse, Tyler Sam, Huey Sun, George Montañez
    Learning and Optimization Conference (LOD), 2020.
    PDF BibTex   Student authors contributed equally

  6. ICAART
    Decomposable Probability-of-Success Metrics in Algorithmic Search
    Tyler Sam, Jake Williams, Abel Tadesse, Huey Sun, George Montañez
    International Conference on Agents and Artificial Intelligence (ICAART), 2020.
    PDF BibTex   Student authors contributed equally


Powered by Jekyll and Minimal Light theme.