Hyunbin Jin

LLM Safety, Computational Social Science

My research interests lie in LLM safety, focusing on developing methods that enhance the trustworthiness of LLMs for real-world deployment. I am also interested in leveraging LLMs to address interdisciplinary challenges at the intersection of AI and the social sciences.

Education

M.S., Data Science — Seoul National University (Sep. 2023 - Feb. 2026)

Exchange Student — New York University (Sep. 2021 - Dec. 2021)

B.A., Political Science and Diplomacy — Yonsei University (Mar. 2018 - Feb. 2023)

Publications

Conference

"Well, Keep Thinking": Enhancing LLM Reasoning with Adaptive Injection Decoding

Hyunbin Jin*, Je Won Yeom*, Seunghyun Bae*, Taesup Kim

ACL 2025 Findings • arXiv

ATAS: Any-to-Any Self-Distillation for Enhanced Open-Vocabulary Dense Prediction

Juan Yeo*, Soonwoo Cha*, Jiwoo Song*, Hyunbin Jin, Taesup Kim

ICCV 2025 • arXiv

Evaluating Structure-Aware Retrieval and Safety in Statute-Centric Legal QA

Kyubyung Chae, Je Won Yeom, Jeongjae Park, Seunghyun Bae, Ijun Jang, Hyunbin Jin, Jinkwan Jang, Taesup Kim

ACL 2026 Oral • arXiv

Preprint

From Threat to Tool: Leveraging Refusal-Aware Injection Attacks for Safety Alignment

Kyubyung Chae*, Hyunbin Jin*, Taesup Kim

arXiv

What If TSF: A Benchmark for Reframing Forecasting as Scenario-Guided Multimodal Forecasting

Jinkwan Jang*, Hyunbin Jin*, Hyungjin Park, Kyubyung Chae, Taesup Kim

arXiv

Contribution

HyperCLOVA X 8B Omni

NAVER Cloud HyperCLOVA X Team - Contributor

Technical Report • arXiv

* : Equal Contribution

Work Experience

Residency Program, NAVER Cloud — Oct. 2025 - Jul. 2026

- Constructed distillation-based synthetic SFT data to enhance reasoning, and designed a safety evaluation framework assessing LRM reasoning trajectories.

NLP Research Engineer, Twigfarm — Oct. 2022 - Feb. 2023

- Developed an image-guided machine translation (IMT) pipeline for Korean–Japanese webtoon translation.

- Constructed a Korean–English parallel dataset for video-guided machine translation (VMT).

Data Analyst Intern, UpennSolution — Feb. 2021 - Jul. 2021

- Conducted quantitative analysis on real-world data to extract insights into social phenomena.

Research Projects

Clinical Information Extraction — Oct. 2025 - Feb. 2026

Research Collaboration with Seoul National University Bundang Hospital

- Constructed an LLM-based pipeline that extracts and structures key clinical information from epilepsy patient notes, ensuring robustness to heterogeneous physician documentation styles.

Legal RAG System — Mar. 2025 - Aug. 2025

Research Collaboration with Naedam C&C

- Constructed a legal retrieval corpus and an evaluation QA dataset for Korean fire safety law, and developed a domain-specific RAG system validated on this benchmark.

Face Blur Estimation — Jul. 2022 - Sep. 2022

Research Collaboration with Alchera

- Developed and evaluated a regression model estimating blur levels in facial images to filter motion-blurred inputs for reliable face recognition.

Patents

Method and Apparatus for Adaptive Injection Decoding to Prevent Incomplete Reasoning in Large Language Models

Patent No. 10-2025-0165083, South Korea (Patent Pending)

Awards & Honors

Fellowships / Scholarships

Support for Next-Generation Researchers Program

Research Subsidy for Master’s Students, Ministry of Education, South Korea (Jul. 2024 - Jun. 2025)

Merit-based Scholarship

Awarded for academic excellence, Seoul National University (2024)

Jilli Scholarship (Merit-based Scholarship)

Awarded for academic excellence, Yonsei University (2018, 2020)

Awards

Honors

Top 10% in the College of Social Science, Yonsei University (2018, 2020)