Papers by Yongdae Kim

Copied to clipboard

Challenge:	Large language models generate human-like text, raising concerns about their misuse in creating deceptive content.
Approach:	They propose a framework for detecting LLM-generated comments in Korean news and introduce a XDAC framework that leverages explainable AI to uncover distinguishing linguistic patterns at token and character levels.
Outcome:	The proposed framework outperforms existing methods and achieves 98.5% F1 score in detection and 84.3% F1 in attribution.