作者位次 #5/7ACL 2026Poster· Natural Language Processing
S 2 O : Early Stopping for Sparse Attention via Online Permutation
Yu Zhang · Songwei Liu · Chenqian Yan · Linsheng · Beichen Ning · Fangmin Chen · XING WANG
摘要
Attention scales quadratically with sequence length, fundamentally limiting long-context inference.Existing block-granularity sparsification can reduce latency, but coarse blocks impose an intrinsic sparsity ceiling, making further improvements difficult even with carefully engineered designs.We...