Seogyeong Jeong

About Me

Profile Picture

Hi! I am a Ph.D. student at KAIST School of Computing, advised by Prof. Alice Oh. My primary research interest is reasoning in large language models (LLMs), with additional interests in post-training and mechanistic interpretability.

My research begins with understanding how reasoning behaviors arise from the internal representations and learning dynamics of LLMs. In particular, I am interested in how these behaviors vary across different operations, languages, cultures, and values. I use mechanistic interpretability to investigate the internal mechanisms underlying such variation and to identify where and why reasoning failures occur.

Building on this understanding, I aim to gradually move from analysis to intervention: developing model-level methods that improve reasoning through pretraining, post-training, alignment, and test-time control. My long-term goal is to build reliable, interpretable, and controllable LLMs that reason robustly across diverse user contexts.

Education

Publications

  1. Seogyeong Jeong, Jaehui Hwang, Dongyoon Han, Geonmo Gu, Alice Oh, Taekyung Kim. 2026. Beneath the Surface of Chain-of-Thoughts: A Mechanistic Interpretation of Reasoning Operations in LLMs. To appear in The 2026 Conference on Empirical Methods in Natural Language Processing(2026 EMNLP Main Conference). Also to be presented at The Ninth Workshop on Analyzing and Interpreting Neural Networks for NLP (BlackboxNLP 2026).
  2. *Seogyeong Jeong, *Kiwoong Park, Seyoung Song, Eunsu Kim, Ken E. Friedl, Jaeho Kim, Alice Oh. 2026. LoCar: Localization-Aware Evaluation of In-Vehicle Assistants through Fine-Grained Sociolinguistic Control. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track)(ACL-Industry2026)
  3. *Seyoung Song, *Seogyeong Jeong, Eunsu Kim, Jiho Jin, Dongkwan Kim, Jay Shin, Alice Oh. 2025. MUG-Eval: A Proxy Evaluation Framework for Multilingual Generation Capabilities in Any Language. In Findings of the Association for Computational Linguistics: EMNLP 2025 (Long). Also presented at the 5th Workshop on Multilingual Representation Learning (MRL 2025).
  4. *Junyeong Park, *Seogyeong Jeong, *Seyoung Song, Yohan Lee, Alice Oh. 2025. LLM-C3MOD: A Human-LLM Collaborative System for Cross-Cultural Hate Speech Moderation. In NAACL 2025 Workshop - C3NLP(Workshop on Cross-Cultural Considerations in NLP) [Slides] [Poster]

Preprint

  1. Soyeon Park, *Seogyeong Jeong, *Sunwoo Kim, Alice Oh. 2026. The Role of Fine-grained Harm Signals in LLM Safety arXiv preprint
  2. Haeun Yu, Seogyeong Jeong, Siddhesh Pawar, Jisu Shin, Jiho Jin, Junho Myung, Alice Oh, Isabelle Augenstein. 2025. Entangled in Representations: Mechanistic Investigation of Cultural Biases in Large Language Models. arXiv preprint
  3. Paul Röttger, Giuseppe Attanasio, Felix Friedrich, Janis Goldzycher, Alicia Parrish, Rishabh Bhardwaj, Chiara Di Bonaventura, Roman Eng, Gaia El Khoury Geagea, Sujata Goswami, Jieun Han, Dirk Hovy, Seogyeong Jeong, Paloma Jeretič, Flor Miriam Plaza-del-Arco, Donya Rooein, Patrick Schramowski, Anastassia Shaitarova, Xudong Shen, Richard Willats, Andrea Zugarini, Bertie Vidgen. 2025. MSTS: A Multimodal Safety Test Suite for Vision-Language Models. arXiv preprint

Work Experience

Awards & Honors

Teaching Experience

SEOGYEONGJEONG