一作NeurIPS 2026Poster
Attention-Based Sampler for Diffusion Language Models
Yuyan Zhou · Kai Syun Hou · Weiyu Chen · James Kwok
长上下文与架构OpenReview
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。
Yuyan Zhou · Kai Syun Hou · Weiyu Chen · James Kwok
Ziwen Zhang · Weiyu Chen · Yuyan Zhou · Yichen Zhu · James Kwok
Yuyan Zhou · Weiyu Chen · James Kwok
Diffusion-based Large Language Models (dLLMs) are emerging as a powerful alternative to traditional autoregressive models. These models learn to generate text by iteratively denoising masked sequences. In this work, we identify a critical problem in dLLMs: the model’s attention is wastefully expe...