一作ICLR 2026Poster· Optimization / Sampling and Optimization
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。
Jiayi Xin · Xiang Li · Evan Qiang · Weiqing He · Tianqi Shang · Weijie J Su · Qi Long
In-context learning (ICL) performance depends critically on which demonstrations are placed in the prompt, yet most existing selectors prioritize heuristic notions of relevance or diversity and provide limited insight into the \textbf{coverage} of a demonstration set. We propose \textbf{Unseen Co...