Showing cs.DBShow all
2 papers · 1 filter
cs.DB2026
GRACE: A Dynamic Coreset Selection Framework for Large Language Model Optimization
Tianhao Tang, Haoyang Li, Lei Chen
Large Language Models (LLMs) have demonstrated remarkable capabilities in natural language understanding and generation. However, their immense number of parameters and complex tra…
cs.DB2026
QCFuse: Query-Centric Cache Fusion for Efficient RAG Inference
Jianxin Yan, Zeheng Qian, Wangze Ni +6
Cache fusion accelerates generation process of LLMs equipped with RAG through KV caching and selective token recomputation, thereby reducing computational costs and improving effic…