Measuring and Mitigating Post-Hoc Rationalization in Reverse Chain-of-Thought Generation
Guangyue Peng · Zongchao Chen · Wen Luo · Yuntao Wen · Wei Li · Ruixiang Feng · Ran Le · Chen Yang · Zhenwei An · Yang Song · Tao Zhang · Houfeng 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。
Guangyue Peng · Zongchao Chen · Wen Luo · Yuntao Wen · Wei Li · Ruixiang Feng · Ran Le · Chen Yang · Zhenwei An · Yang Song · Tao Zhang · Houfeng Wang
Ruixiang Feng · Yuntao Wen · Silin Zhou · Ke Shi · Yifan Wang · Ran Le · Zhenwei An · Zongchao Chen · Chen Yang · Guangyue Peng · Yiming Jia · Dongsheng Wang · Tao Zhang · Lisi Chen · Yang Song · Shen Gao · Shuo Shang
Language Reasoning Models (LRMs) achieve strong performance by scaling test-time computation but often suffer from “overthinking”, producing excessively long reasoning traces that increase latency and memory usage. Existing LRMs typically enforce conciseness with uniform length penalties, which o...