1 paper · 1 filter
Yijie Hao, Lingjie Chen, Ali Emami +1
Can we trust the reasoning traces that large reasoning models (LRMs) produce? We investigate whether these traces faithfully reflect what drives model outputs, and whether models w…