Publications

2025

1. Memory

   

   Embodied Agents Meet Personalization: Exploring Memory Utilization for Personalized Assistance

   Taeyoon Kwon^†,  Dongwook Choi^†,  Sunghwan Kim ,  Hyojun Kim,  Seungjun Moon,  Beong-woo Kwak,  Kuan-Hao Huang, and  Jinyoung Yeo

   Arxiv preprint

   arXiv
2. Reward Model

   

   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, and 1 more author

   Arxiv preprint

   arXiv
3. Reward Model

   

   Rethinking Reward Model Evaluation Through the Lens of Reward Overoptimization

   Sunghwan Kim^† ,  Dongjin Kang^†,  Taeyoon Kwon,  Hyungjoo Chae,  Dongha Lee, and  Jinyoung Yeo

   ACL 2025

   arXiv
4. Interaction

   

   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, and  Jinyoung Yeo

   ICLR 2025

   arXiv Code

2024

1. Reward Model

   

   Evaluating Robustness of Reward Models for Mathematical Reasoning

   Sunghwan Kim^† ,  Dongjin Kang^†,  Taeyoon Kwon,  Hyungjoo Chae,  Jungsoo Won,  Dongha Lee, and  Jinyoung Yeo

   Arxiv preprint

   arXiv Code
2. Dialogue

   

   Cactus: Towards Psychological Counseling Conversations using Cognitive Behavioral Theory

   Suyeon Lee^†,  Sunghwan Kim^† ,  Minju Kim^†,  Dongjin Kang,  Dongil Yang,  Harim Kim,  Minseok Kang,  Dayi Jung,  Min Hee Kim,  Seungbeen Lee,  Kyoung-Mee Chung,  Youngjae Yu,  Dongha Lee, and  Jinyoung Yeo

   EMNLP 2024 findings

   arXiv Code
3. Reasoning

   

   Language Models as Compilers: Simulating Pseudocode Execution Improves Algorithmic Reasoning in Language Models

   Hyungjoo Chae,  Yeonghyeon Kim,  Seungone Kim,  Kai Tzu-iunn Ong,  Beong-woo Kwak,  Moohyeon Kim,  Sunghwan Kim ,  Taeyoon Kwon,  Jiwan Chung,  Youngjae Yu, and  Jinyoung Yeo

   EMNLP 2024

   arXiv Code
4. Dialogue

   

   Can Large Language Models be Good Emotional Supporter? Mitigating Preference Bias on Emotional Support Conversation

   Dongjin Kang^†,  Sunghwan Kim^† ,  Taeyoon Kwon,  Seungjun Moon,  Hyunsouk Cho,  Youngjae Yu,  Dongha Lee, and  Jinyoung Yeo

   ACL 2024

   🏆 Outstanding Paper Award 🏆

   Abs arXiv

   Outstanding Paper Award

   Emotional Support Conversation (ESC) is a task aimed at alleviating individuals’ emotional distress through daily conversation. Given its inherent complexity and non-intuitive nature, ESConv dataset incorporates support strategies to facilitate the generation of appropriate responses. Recently, despite the remarkable conversational ability of large language models (LLMs), previous studies have suggested that they often struggle with providing useful emotional support. Hence, this work initially analyzes the results of LLMs on ESConv, revealing challenges in selecting the correct strategy and a notable preference for a specific strategy. Motivated by these, we explore the impact of the inherent preference in LLMs on providing emotional support, and consequently, we observe that exhibiting high preference for specific strategies hinders effective emotional support, aggravating its robustness in predicting the appropriate strategy. Moreover, we conduct a methodological study to offer insights into the necessary approaches for LLMs to serve as proficient emotional supporters. Our findings emphasize that (1) low preference for specific strategies hinders the progress of emotional support, (2) external assistance helps reduce preference bias, and (3) existing LLMs alone cannot become good emotional supporters. These insights suggest promising avenues for future research to enhance the emotional intelligence of LLMs.