一作ICML 2026Poster· Social Aspects / Alignment
Real-Time Monitoring and Calibration of Chain-of-Thought Sycophancy in Large Reasoning Models
Jingyu Hu · Shu Yang · Xilin Gong · Hongming Wang · Weiru Liu · Di Wang
首次获取该候选人的完整档案及 StartLux 评估共 20 积分;每位用户单独解锁,之后查看免费。主动刷新 20 积分。
实时查找电子邮箱,费用不超过20积分;未找到邮箱不扣费。
Jingyu Hu · Shu Yang · Xilin Gong · Hongming Wang · Weiru Liu · Di Wang
Shu Yang · Jingyu Hu · Tong Li · Hanqi Yan · Wenxuan Wang · Di Wang
We introduce AutoMonitor-Bench, the first benchmark designed to systematically evaluate the reliability of LLM-based misbehavior monitors across diverse tasks and failure modes. AutoMonitor-Bench consists of 3,010 carefully annotated test samples spanning question answering, code generation, and...
Jingyu Hu · Mengyue Yang · Mengnan Du · Weiru Liu