Jungsoo Park
Ph.D. Student, Georgia Tech
School of Interactive Computing
Coda S1153A
jpark3272 [AT] gatech.edu
About
Hello! I am a Ph.D. student in the School of Interactive Computing at Georgia Tech, advised by Alan Ritter. Before starting my Ph.D., I worked as an Applied Scientist at NAVER Clova, where I contributed to the development of a vision-language foundation model, HyperCLOVA X Vision. I received my B.E. in Statistics, B.S. in Computer Science & Engineering, and M.E. in Software from Korea University.
My research focuses on improving the reasoning and prediction capabilities of language models by learning from distributions of model trajectories. I am interested in developing methods to robustly optimize over these trajectories and extract useful signals from them to improve model behavior. I am particularly interested in test-time discovery, and more broadly in connecting it with LLM-based prediction to enable more generalizable discovery and decision-making.
Education
Georgia Institute of Technology Atlanta, GA, USA
Ph.D. in Computer Science Aug. 2024 - Present
Advisor: Alan Ritter
Korea University Seoul, Korea
M.E. in Computer Science & Engineering Mar. 2019 - Aug. 2021
Advisor: Jaewoo Kang
Korea University Seoul, Korea
B.E. in Statistics, B.S. in Computer Science & Engineering Mar. 2013 - Feb. 2019
Magna Cum Laude
Research Experience
Allen Institute of Artificial Intelligence Seattle, WA, USA
Research Intern (Ph.D.) May. 2026 - Present
Mentor: Jay DeYoung
Georgia Institute of Technology Atlanta, GA, USA
Graduate Research Assistant (Ph.D.) Aug. 2024 - Present
Advisor: Alan Ritter
Naver Clova Seongnam, Korea
Applied Scientist Jul. 2021 - Aug. 2024
Research Intern July. 2020 - Dec. 2020
University of Washington Seattle, WA, USA
Visiting Scholar (remote) Jan. 2021 - Jul. 2021
Mentor: Sewon Min Advisor: Hannaneh Hajishirzi, Luke Zettlemoyer
Korea University Seoul, Korea
Graduate Research Assistant (M.E.) Mar. 2019 - Aug. 2021
Advisor: Jaewoo Kang
Publications
* indicates equal contribution.Distribution-Aware Reward: Reinforcement Learning over Predictive Distributions for LLM Regression
Jungsoo Park, Hyungjoo Chae, Ethan Mendes, Jay DeYoung, Varsha Kishore, Wei Xu, Alan Ritter
arXiv preprint
Soft Token Alignment for Cross-Lingual Reasoning
Ivy He, Jungsoo Park, Wei Xu, Alan Ritter
NeurIPS 2026
Who Pays More for Safety? Measuring the Disparate Cost of Safety Alignment across Languages
Chanwoong Yoon, Jungsoo Park, Alan Ritter
EMNLP 2026
Safe and Scalable Web Agent Learning via Recreated Websites
Hyungjoo Chae, Jungsoo Park, Alan Ritter
ICML 2026
Didactic to Constructive: Turning Expert Solutions into Learnable Reasoning
Ethan Mendes, Jungsoo Park, Alan Ritter
ICML 2026
Anticipatory Evaluation of Language Models
Jungsoo Park, Ethan Mendes, Gabriel Stanovsky, Alan Ritter
arXiv preprint
Can LLMs Help Uncover Insights about LLMs? A Large-Scale, Evolving Literature Analysis of Frontier LLMs
Jungsoo Park, Junmo Kang, Gabriel Stanovsky, Alan Ritter
ACL 2025
Optimizing Test-Time Query Representations for Dense Retrieval
Mujeen Sung, Jungsoo Park, Jaewoo Kang, Danqi Chen, Jinhyuk Lee
EMNLP-Findings 2023
Empowering Sentence Encoders with Prompting and Label Retrieval for Zero-shot Text Classification
Jimin Hong*, Jungsoo Park*, Daeyoung Kim*, Seongjae Choi, Bokyung Son, Jaewook Kang
arXiv preprint
FaVIQ: FAct Verification from Information-seeking Questions
Jungsoo Park*, Sewon Min*, Jaewoo Kang, Luke Zettlemoyer, Hannaneh Hajishirzi
ACL 2022
Consistency Training with Virtual Adversarial Discrete Perturbation
Jungsoo Park*, Gyuwan Kim*, Jaewoo Kang
NAACL 2022
Learn to Resolve Conversational Dependency: A Consistency Training Framework for Conversational Question Answering
Gangwoo Kim, Hyunjae Kim, Jungsoo Park, Jaewoo Kang
ACL 2021
Adversarial Subword Regularization for Robust Neural Machine Translation
Jungsoo Park, Mujeen Sung, Jinhyuk Lee, Jaewoo Kang
EMNLP Findings 2020