eRLT: Efficient VLA Reinforcement Learning via Action-Relevant Token Routing

作者：Dehao Huang, Jianbang Liu, Jianpan Gao, Chao Tang, Zilang Cen, Zedong Dan, Jiaheng Wang, Tingguang Li, Yue Wang, Hong Zhang · 单位：Southern University of Science and Technology, Shenzhen, China., Beijing Zhongguancun Academy, Beijing, China., Wuhan University, Wuhan, China., Sun Yat-sen University, Guangzhou, China. · 会议/期刊：arXiv preprint · 方向：cs.RO · 发布日期：2026-10-02