Sue Hyun Park

prof_pic.jpg

Hello! I am a second year M.S. student at KAIST AI. I am fortunate to be advised by Minjoon Seo in Language & Knowledge Lab. Previously, I received a B.B.A. and a B.S. degree from Seoul National University. I also interned at FriendliAI in its early stage, envisioning the business impact of generative AI models.

My research lies in the area of natural language processing and machine learning. I study enhancing generalization of large (vision) language models using a data-centric approach and understanding complex reasoning abilities in large language models. My research interests include pluralistic alignment, inductive and causal reasoning, and AI safety and security. Here is my CV.

news

Jul 01, 2024 Our new preprint, Hierarchical Deconstruction of LLM Reasoning: A Graph-Based Framework for Analyzing Knowledge Utilization (previous title: Investigating How Large Language Models Leverage Internal Knowledge to Perform Complex Reasoning), is out. (Update Sep 21, 2024: Accepted to EMNLP 2024!)
Jun 30, 2024 Our new preprint, BiGGen-Bench, is out. I contributed to creating fun Theory of Mind tasks like the time traveler’s dilemma!
May 31, 2024 Our new preprint, Aligning to Thousands of Preferences via System Message Generalization, is out. (Update July 17, 2024: Accepted to ACL 2024 Workshop NLP4ConvAI for oral presentation! Update Sep 26, 2024: Accepted to NeurIPS 2024!)
Jan 12, 2024 Our new preprint, Prometheus-Vision, is out. (Update May 16, 2024: Accepted to ACL 2024 Findings!)
Nov 14, 2023 Our new preprint, Volcano, is out. (Update Mar 14, 2024: Accepted to NAACL 2024!)
Aug 28, 2023 I started my M.S. program at KAIST AI.

selected papers

  1. Aligning to Thousands of Preferences via System Message Generalization
    Seongyun Lee* , Sue Hyun Park* , Seungone Kim , Minjoon Seo
    NeurIPS 2024
  2. Hierarchical Deconstruction of LLM Reasoning: A Graph-Based Framework for Analyzing Knowledge Utilization
    Miyoung Ko* , Sue Hyun Park* , Joonsuk Park† , Minjoon Seo†
    EMNLP 2024
  3. Volcano: Mitigating Multimodal Hallucination through Self-Feedback Guided Revision
    Seongyun Lee , Sue Hyun Park , Yongrae Jo , Minjoon Seo
    NAACL 2024

invited talks

Feb 16, 2024 NAVER Tech Meetup
Title: Fine-grained Evaluation of Vision-Language Models through “Vision-Language Model as a Judge”

services

Secondary Conference Reviewer ACL 2023, EMNLP Industry Track 2023

education

Aug 2023 - Aug 2025 (expected) Kim Jaechul Graduate School of AI, KAIST (KAIST AI)
Master of Science in Artificial Intelligence, advised by Minjoon Seo
Mar 2018 - Aug 2023 Seoul National University
Bachelor of Business Administration
Bachelor of Science in Artificial Intelligence (Interdisciplinary major)

work experience

Dec 2022 - Aug 2023 Language & Knowledge Lab, KAIST AI
Undergraduate research assistant
Jul 2022 - Dec 2022 Software Platform Lab, Seoul National University
Undergraduate research assistant
Mar 2021 - Oct 2022 Artificial Intelligence Institute of Seoul National University
Technical writer [blog]
Feb 2021 - Dec 2021 FriendliAI
Corporate development & Business development intern, worked closely with Byung-gon Chun