Hyun Ryu 류현

Master’s student · Graduate School of AI, KAIST

Hi! I’m Hyun, a master’s student at KAIST AI, advised by Prof. Eunho Yang. I study how to make large language models better at reasoning, argumentation, and critical thinking.

My research focuses on model training and evaluation on non-verifiable tasks, including mid-training for improving critical thinking ability (Arguinas), rubric-based RL training (FORK), and paper review evaluation (ReviewScore). I also worked on neuro-symbolic approaches, including natural language formalization (CLOVER, ReviewScore, Arguinas).

Previously, I was a visiting student at Carnegie Mellon University’s Language Technologies Institute, co-advised by Carolyn Rosé and Sean Welleck. I interned at LG AI Research working with Moontae Lee, and I also collaborated with Gregor Betz at Karlsruhe Institute of Technology’s Department of Philosophy. I received my B.S. in Electrical Engineering from KAIST.