作者位次 #2/7ICLR 2026Poster
Thinking-Free Policy Initialization Makes Distilled Reasoning Models More Effective and Efficient Reasoners
Xin Xu · Clive Bai · Kai Yang · Tianhao Chen · Yang Wang · Saiyong Yang · Can Yang
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。
Xin Xu · Clive Bai · Kai Yang · Tianhao Chen · Yang Wang · Saiyong Yang · Can Yang
Chenming Tang · Hsiu-Yuan Huang · Weijie Liu · Clive Bai · Saiyong Yang · Yunfang Wu
Reinforcement learning with verifiable rewards (RLVR) has significantly boosted the reasoning capability of language models (LMs). However, existing RLVR approaches train LMs based on their own on-policy responses and are constrained by the initial capability of LMs, thus prone to exploration sta...
Hsiu-Yuan Huang · Chenming Tang · Weijie Liu · Clive Bai · Saiyong Yang · Yunfang Wu
Existing Reinforcement Learning from Verifiable Rewards (RLVR) methods, such as Group Relative Policy Optimization (GRPO), have achieved remarkable progress in improving the reasoning capabilities of Large Reasoning Models (LRMs). However, they exhibit limited exploration due to reliance on on-po...