一作ICML 2026Poster· Deep Learning / Large Language Models
One-shot Entropy Minimization for Language Model Reasoning
Zitian Gao · Yilong Chen · Haoming Luo · Joey Zhou · Bryan Dai
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。
Zitian Gao · Yilong Chen · Haoming Luo · Joey Zhou · Bryan Dai
Yilong Chen · Zitian Gao · Yihao Xiao · Jason Klein Liu · Xinyu Yang · Yifan Luo · Haoming Luo · Zhengmao Ye · Tingwen Liu · Ran Tao · Bryan Dai
Although the Universal Transformer (UT) mitigates the diminishing returns of standard LLM scaling by decoupling parameter count from depth, it remains constrained by linear computational costs and rigid weight-sharing mechanisms. These limitations lead to severe functional homogeneity, which subs...