Research Highlights

  • arXiv'26 / Sequential Behavioral Watermarking for LLM Agents / GitHub / 요약
  • arXiv'26 / Adaptive Steering and Remasking for Safe Generation in Diffusion Language Models / GitHub / 요약
  • arXiv'26 / Cross-Family Universality of Behavioral Axes via Anchor-Projected Representations / GitHub / 요약
  • arXiv'26 / NCO: A Versatile Plug-in for Handling Negative Constraints in Decoding / GitHub / 요약
  • arXiv'26 / From Intuition to Expertise: Rubric-Based Cognitive Calibration for Human Detection of LLM-Generated Korean Text / GitHub / 요약
  • ACL'26 / A Linguistics-Aware LLM Watermarking via Syntactic Predictability / GitHub / 요약
  • ACL'26 / RV-HATE: Reinforce Multi-Module Voting for Implicit Hate Speech Detection / GitHub / 요약
  • ACL-Findings'26 / ContractEval: A Benchmark for Evaluating Contract-Satisfying Assertions in Code Generation / GitHub / 요약
  • arXiv'26 / How Does the Thinking Step Influence Model Safety? An Entropy-based Safety Reminder for LRMs / GitHub / 요약
  • arXiv'26 / KOTOX: A Korean Toxic Dataset for Deobfuscation and Detoxification / GitHub / 요약
  • arXiv'26 / Steering Language Models Before They Speak: Logit-Level Interventions
  • AAAI'26 / WaterMod: Modular Token-Rank Partitioning for Probability-Balanced LLM Watermarking / GitHub / 요약
  • EACL'26 / Reparing Regex Vulnerabilities via Localization-Guided Instructions / GitHub / 요약
  • EACL'26 / DITTO: A Spoofing Attack Framework on Watermarked LLMs via Knowledge Distillation / GitHub / 요약
  • arXiv'25 / RegexPSPACE: A Benchmark for Evaluating LLM Reasoning on PSPACE-Complete Regex Problems / GitHub / summary / 요약
  • arXiv'25 / ECO: Enhanced Code Optimization via Performance-Aware Prompting / GitHub / 요약
  • arXiv'25 / MEC^3O: Multi-Expert Consensus for Code Time Complexity Prediction / GitHub / 요약
  • EACL-Findings'26 / Marking Code Without Breaking It: Code Watermarking for Detecting LLM-Generated Code / GitHub / 요약


    NEWS

  • 연구실에 지원할 학생들은 모집주제 페이지를 먼저 확인해주세요.
  • We have two papers accepted for CIAA 2026; one is about the pattern mining and the other is about the decomposition of regular languages. The second one is joint work with Prof. Kai Salomaa from Queen's University in Canada.
  • Our "ReSyn: A Generalized Recursive Regular Expression Synthesis Framework" paper is accepted for IJCAI 2026. Good job, Su-Hyeon. This is joint work with Prof. Sang-Ki Ko's group from University of Seoul.
  • We have three papers accepted for ACL 2026; two for main (LLM watermarking and hate dection) and one for findings (code benchmark). Congrats to all co-authors!
  • We have five papers accepted for EACL 2026; three for main, two for findings. Congrats to all co-authors! Especially, Suyeon---a co-author of our DITTO paper--- was an undergrad intern in our lab, GREAT.
  • Our "WaterMod: Modular Token-Rank Partitioning for Probability-Balanced LLM Watermarking " paper is accepted for AAAI 2026. Good job, Shinwoo and co-authors. This is joint work with Hyejin from Rensselaer Polytechnic Institute.
  • We have four papers accepted for EMNLP 2025; two for main, two for findings. Congrats to all co-authors!
  • We have one paper (Pattern Mining under Simon's Congruence) accepted for DLT 2025, which will be held in Korea for the first time! Congrats, Sungmin.
  • We have one paper (KatFishNet: Detecting LLM-Generated Korean Text through Linguistic Feature Analysis) accpeted for ACL 2025. Well done, Shinwoo, Shubin, Do-Kyung---especially, Shubin was an undergrad intern in our lab, AMAZING!
  • Our "LogiCase: Effective Test Case Generation from Logical Description in Competitive Programming" paper is accepted for IJCAI 2025. Good job, Sicheol. This is joint work with Prof. Sang-Ki Ko's group from University of Seoul.
  • Our "TCProF: Time-Complexity Prediction SSL Framework" paper is accepted for NAACL 2025. A more detail info will be updated soon. Congrats Joonghyuk, Hyeseon, Jungin and Soohan.
  • Our "Impact of Large Language Models of Code on Fault Localization" paper is accepted for ICST 2025. This is joint work with Prof. Hyeonseung Im's group from Kangwon National University.
  • Our "Analyzing Offensive Language Dataset Insights from Training Dynamics and Human Agreement Level" paper is accepted for COLING 2025. Congrats Do-Kyung and Hyeseon.


    RESEARCH INTERESTS


    Contact Information
    School of Computing, Yonsei University
    50 Yonsei-Ro, Seodaemun-Gu, Seoul 03722, Republic of Korea
    Email: emmous [at] yonsei [dot] ac [dot] kr

    Last updated 2026/4/7