Wonje Jeung

MS student, Yonsei University

1 paper at NeurIPS 2025

Homepage· OpenReview· Semantic Scholar· Google Scholar

Poster Session 2

1 paper

Wednesday, December 3, 2025 · 4:30 PM → 7:30 PM

Exhibit Hall C,D,E

SAFEPATH: Preventing Harmful Reasoning in Chain-of-Thought via Early Alignment

#1405 · Wonje Jeung, Sangyeon Yoon, Minsuk Kahng, Albert No

We propose SAFEPATH, a lightweight method that aligns Large Reasoning Models to detect and suppress harmful chain-of-thought reasoning by injecting a brief safety signal at the start of reasoning.