About

Hi! I'm Hyeonbin.

I’m a Master’s student at KAIST AI advised by Minjoon Seo.

My graduate research began with a simple question: How do language models reason, and how can we make them reason better? Along the way, I quickly ran into what Sutton calls the "bitter lesson": models often appear capable by memorizing patterns in the training data, only to plateau once they encounter situations outside that distribution.

Nevertheless, I study how to design learning signals that encourage models to move past surface-level pattern matching, and I aim to explore these capabilities through diverse domains. Currently, I am an intern at ALIN Lab, where I work on robotics + reinforcement learning.

I’m always open to collaborations—feel free to reach out!

Send Mail View Publications