1 paper
Yinsheng Yao, Jiehao Tang, Zhaozhen Yang +1
While explicit reasoning trajectories enhance model interpretability, existing paradigms often rely on monolithic chains that lack intermediate verification, allowing early errors…