Unverified50% confidenceFactExact time
RadixAttention 引入全局共享 KV Cache 池，通过引用计数和 LRU 淘汰策略管理缓存生命周期，并在 Radix Tree 中执行最长前缀匹配
1
Sources
50%
Confidence
Long-term
Relevance
7/16/2026
First Seen
Sources
Related Claims
UnverifiedRadixAttention 通过将 KV Cache 组织为前缀树结构，使具有相同前缀的多个请求共享同一份缓存75% similarUnverifiedSGLang's RadixAttention technology uses a radix tree structure to cache shared prompt prefixes across different requests, so when 100 users use the same system prompt, the KV Cache only needs to be computed once.63% similarUnverifiedIn scientific research scenarios, RAG is typically used in combination with vector databases such as Pinecone and Weaviate.61% similarUnverifiedSGLang's core features include RadixAttention (KV cache sharing), Continuous Batching, and native structured output constraints.61% similarUnverifiedIn the offline phase of the traditional RAG process, documents are chunked and converted into vectors via an embedding model and stored in a vector database; in the online phase, similar document chunks are retrieved using an Approximate Nearest Neighbor (ANN) algorithm.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/533861API
curl https://kongchang.com/api/v1/knowledge/claims/533861MCP
get_claim(id=533861)