Unverified50% confidenceFactExact time
The bottleneck of single-pass retrieval stems from vector similarity search compressing the user query into a single high-dimensional vector; when a question contains multiple semantic directions, this vector becomes an average of multiple semantics, causing sub-intents to be insufficiently represented.
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedRelying solely on vector similarity retrieval often yields imprecise recall, necessitating a hybrid retrieval approach that combines keyword-based retrieval.78% similarUnverifiedMulti-hop queries need to traverse multiple information nodes in a step-by-step manner, and vector retrieval is ill-suited for such scenarios due to its lack of a relational dimension.77% similarUnverifiedRAG suffers from a scattered-answer problem because the current retrieval approach scores relevance between a query and individual chunks, making it difficult to guarantee that multiple complementary chunks are retrieved simultaneously.76% similarUnverifiedThe shortcomings of Naive RAG include: single vector retrieval tends to overlook exact keyword matching, there is no metadata filtering, and the lack of reranking results in insufficiently refined relevance ordering.72% similarUnverifiedThe frequency with which web search is triggered varies enormously across platforms and models, and triggering search more frequently does not necessarily lead to higher-quality answers.72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/977278API
curl https://kongchang.com/api/v1/knowledge/claims/977278MCP
get_claim(id=977278)