HHR: Hierarchical Hash Retrieval for Efficient LLM Generation
H Hierarchical Hash Retrieval (HHR) is proposed, a coarse-to-fine framework that progressively improves retrieval accuracy through Geometry-Aware Key Routing and Learned Hash Projection, substantially improving the fidelity of hash-based sparse attention.
Lian Liu, Tian-Tian Zheng, You Huang et al.
· 0 citations