Unverified50% confidenceFactExact time
Sparse vector retrieval typically uses an inverted index, where each term maintains a posting list recording the document IDs containing that term and their weights.
1
Sources
50%
Confidence
Long-term
Relevance
10/20/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedClassification should be set when the agent closes the ticket, in at most two clicks; time-consuming classification becomes sloppy, and sloppy classification is worse than none because it looks like data.69% similarUnverifiedTo improve cache hit rates, historical bytes in long-running tasks should follow an append-only discipline, and IDs for the same file should use deterministic suffixes rather than random IDs.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1075143API
curl https://kongchang.com/api/v1/knowledge/claims/1075143MCP
get_claim(id=1075143)