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。
Yuzhong Zhao · Yue Liu · Junpeng Liu · Jingye Chen · xun wu · Yaru Hao · Tengchao Lv · Shaohan Huang · Lei Cui · Qixiang Ye · Fang Wan · Furu Wei
Siwei Wang · Yifei Shen · Haoran Sun · Shi Feng · Shang-Hua Teng · Li Dong · Yaru Hao · Wei Chen
Di Zhang · xun wu · Shaohan Huang · Lingjie Jiang · Yaru Hao · Li Dong · Zewen Chi · Zhifang Sui · Furu Wei
Reinforcement learning with verifiable rewards (RLVR) has emerged as a powerful paradigm for improving reasoning capabilities. However, training RLVR with Mixture-of-Experts (MoE) policies remains fragile and is often prone to reward collapse.We identify a MoE-specific source of instability, refe...