ParisKV: Fast and Drift-Robust KV-Cache Retrieval for Long-Context LLMs 文章

ArXiv CS.CL2026-06-01NEWSen作者: Yanlin Qi, Xinhang Chen, Huiqiang Jiang, Qitong Wang, Botao Peng, Themis Palpanas

查看原文 →

ParisKV: Fast and Drift-Robust KV-Cache Retrieval for Long-Context LLMs · 相关技术

相关技术

full attention quantized inner-product reranking estimator collision-based candidate selection ParisKV KV-cache retrieval Unified Virtual Addressing (UVA)