Paper Accepted to Findings of EMNLP 2026
Our paper, “Do Backdoored LLMs Share Internal Trigger Representations? Evidence from Frozen SAE Feature Banks,” by Minjeong Choi, Jaesin Ahn, and Heechul Jung, has been accepted to…
SAIL Laboratory Website Is Now Online
The Safe & Applied Intelligence Lab website is now online. This site will document our research, publications, people, lectures, challenges, and collaboration opportunities as the…