Minju Gwak
M.S. Student in Artificial Intelligence at Yonsei University
mjgwak@yonsei.ac.kr
About
Minju Gwak is a graduate researcher at Yonsei University studying how large language models reason. Her work examines the cognitive and information-theoretic structure of LLM reasoning—how models generate, revise, and validate multi-step thoughts in mathematical and symbolic tasks. She develops process-level analyses that track entropy dynamics, reasoning traces, and internal signals that shape model certainty and reflection. Her research also explores reasoning-driven GUI agents and world models that integrate symbolic structure with neural representations. She aims to build interpretable, cognition-aware LLMs whose reasoning processes are as transparent and reliable as their outputs. She is advised by Prof. Jaehyung Kim in the Department of Artificial Intelligence at Yonsei University.
News
2025
Revisiting the Uniform Information Density Hypothesis in LLM Reasoning Traces
Minju Gwak, Guijin Son, Jaehyung Kim
The 5th Workshop on Mathematical Reasoning and AI, NeurIPS 2025 / Under Review at ACL 2026
Web-Shepherd: Advancing PRMs for Reinforcing Web Agents
Hyungjoo Chae*, Sunghwan Kim*, Junhee Cho*, Seungone Kim, Seungjun Moon, Gyeom Hwangbo, Dongha Lim, Minjin Kim, Yeonjun Hwang, Minju Gwak, Dongwook Choi, Minseok Kang, Gwanhoon Im, ByeongUng Cho, Hyojun Kim, Jun Hee Han, Taeyoon Kwon, Minju Kim, Beong-woo Kwak, Dongjin Kang, Jinyoung Yeo
NeurIPS 2025 (Spotlight)
Can You Share Your Story? Modeling Clients' Metacognition and Openness for LLM Therapist Evaluation
Minju Kim, Dongje Yoo, Yeonjun Hwang, Minseok Kang, Namyoung Kim, Minju Gwak, Beong-woo Kwak, Hyungjoo Chae, Harim Kim, Yunjoong Lee, Min Hee Kim, Dayi Jung, Kyong-Mee Chung, Jinyoung Yeo
ACL 2025 Findings
Towards Lifelong Dialogue Agents via Relation-aware Memory Construction and Timeline-augmented Response Generation
Kai Tzu-iunn Ong, Namyoung Kim, Minju Gwak, Hyungjoo Chae, Taeyoon Kwon, Yohan Jo, Seung-won Hwang, Dongha Lee, Jinyoung Yeo
NAACL 2025
Web Agents With World Models: Learning and Leveraging Environment Dynamics in Web Navigation
Hyungjoo Chae, Namyoung Kim, Kai Tzu-iunn Ong, Minju Gwak, Gwanwoo Song, Jihoon Kim, Sunghwan Kim, Dongha Lee, Jinyoung Yeo
ICLR 2025 / NeurIPS 2024 Sys2Reasonig at Scale Workshop
( * indicates equal contribution )
-
OnelineAI Oct. 2025 - PresentResearch Intern
Advisor: Guijin Son • Developed a benchmark for reasoning model evaluation -
Yonsei University 2025 - PresentM.S. in Artificial Intelligence
Advisor: Jaehyung Kim • Research Areas: Large Reasoning Models -
UIC, Yonsei University 2017 - 2024B.A. Justice and Civil Leadership (pre-law) / B.S. Creative Technology Management
Graduation Project: Developing and designing an IoT product for work-from-home workers