Author name: Ziyang Wang, Jiangfeng Xiao, Chuan Xiao, Ruoxiang Li, Rui Mao, Jianbin Qin

arXiv:2604.19398v1 Announce Type: new
Abstract: Large language models (LLMs) are expensive to serve because model parameters, attention computation, and KV caches impose substantial memory and latency costs. We present GRASPrune, a structured pruning …