Attention as a Compass: Efficient Exploration for Process-Supervised RL in Reasoning Models

Runze Liu, Kun Gai, Fuzheng Zhang, Zhihui Xie, Yuling Shi(Tongji University), Lei Lin(Purdue University West Lafayette), Jiakang Wang, Xiaodong Gu(Fudan University), Jian Zhao(Chinese Research Academy of Environmental Sciences), Xiu Li(Xinjiang Production and Construction Corps), Kaiyan Zhang, Guorui Zhou, Chenxin An, Wenping Hu(University of Science and Technology of China)
arXiv (Cornell University)
September 30, 2025
Cited by 0


Related Papers