Dongkeun Yoon

About Me

Hello! I am an MS & PhD integrated student at KAIST Language and Knowledge Lab. I am interested in second language adaptation of language models. More specifically, I want to convert strong language models such as LLaMa to low-resource languages (LangBridge). I recently started a project focused on the code generation capabilities of LLMs and am eager to discuss this further. Please feel free to reach out to me!

Research Interests

Multilingualism, Code Generation, Large Language Models

News

[Mar 2024] LangBridge was accepted to ICLR 2024 Me-FoMo workshop.
[June 2023] I was accepted to KAIST MS & PhD integrated program!
[May 2023] Our Gradient Ascent Post-training and Knowledge Unlearning papers were accepted to ACL 2023 main conference.

Publications

2024

LangBridge: Multilingual Reasoning Without Multilingual Supervision

[Dongkeun Yoon], Joel Jang, Sungdong Kim, Seungone Kim, Sheikh Shafayat, Minjoon Seo

Arxiv preprint

Code

2023

Gradient Ascent Post-training Enhances Language Model Generalization

[Dongkeun Yoon*], Joel Jang*, Sungdong Kim, Minjoon Seo (*Equal Contribution)

ACL 2023

Code

Knowledge Unlearning for Mitigating Privacy Risks in Language Models

Joel Jang, [Dongkeun Yoon], Sohee Yang, Sungmin Cha, Moontae Lee, Lajanugen Logeswaran, Minjoon Seo

ACL 2023

Code

2022

Improving Graph-based Document-Level Relation Extraction Model with Novel Graph Structure

Seongsik Park, [Dongkeun Yoon], Harksoo Kim

CIKM 2022

Teaching Experience

Teaching Assistant at KAIST

Large Language Models (AI 705), Spring 2024