Profile picture

PhD Student

Graduate School of Information Sciences
Tohoku University

Sendai, Japan

Research

I’m a third-year PhD student in the Graduate School of Information Sciences at Tohoku University and a member of Fai Lab and the Tohoku NLP Group. I’m supervised by Professor Jun Suzuki.

My research focuses on multilingualism in large language models, particularly cross-lingual generalization, target-language generation, and internal language representations. I am also interested in interpretability and cultural aspects of LLMs.

Background

2023.4–Present:
Tohoku University, Doctoral Program, Graduate School of Information Sciences

Contact

Email: xie.suchun.p7 at_mark dc.tohoku.ac.jp

GitHub: https://github.com/waterluck


Publications

Can Language Neuron Intervention Reduce Non-Target Language Output? [Journal Version]

Suchun Xie, Hwichan Kim, Shota Sasaki, Kosuke Yamada, Jun Suzuki
Journal of Natural Language Processing (自然言語処理), 2026 (accepted).
Extended journal version. Accepted for publication in Vol. 33, No. 4 (December 2026).
Acceptance Announcement

Can Language Neuron Intervention Reduce Non-Target Language Output? [Conference Version]

Suchun Xie, Hwichan Kim, Shota Sasaki, Kosuke Yamada, Jun Suzuki
Proceedings of the 8th BlackboxNLP Workshop: Analyzing and Interpreting Neural Networks for NLP, pp. 452–466, 2025-11.
Conference version published at BlackboxNLP 2025.
PDF · ACL Anthology

A Multimodal Dialogue System to Lead Consensus Building with Emotion-Displaying

Shinnosuke Nozue, Yuto Nakano, Shoji Moriya, Tomoki Ariyama, Kazuma Kokuta, Suchun Xie, Kai Sato, Shusaku Sone, Ryohei Kamei, Reina Akama, Yuichiroh Matsubayashi, Keisuke Sakaguchi
Proceedings of the 25th Annual Meeting of the Special Interest Group on Discourse and Dialogue, pp. 669–673, 2024-09.
PDF · ACL Anthology

Error Correction on Digitization of Modern Japanese Documents with Pre-trained BERT

Suchun Xie, Akiyo Matsumoto
Proceedings of the 29th Annual Meeting of the Association for Natural Language Processing, pp. 1616–1620, 2023-03.
PDF · Conference Website