KV-Latent: Dimensional-level KV Cache Reduction with Frequency-aware Rotary Positional Embedding

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Shi, Luohe, Li, Zuchao, Zhang, Lefei, Liu, Guoming, Qi, Baoyuan, Zhao, Hai
Format:	Preprint
Published:	2025
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2507.11273