Unverified50% confidenceFactExact time
When chunks are too small, they lose context; the retrieved passage reads fine on its own, but the LLM cannot answer, because the answer's basis falls in other chunks that were not recalled.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedRe-chunking and re-embedding an entire document wastes compute and loses the continuous identity of the content, making it impossible to answer questions about the historical versions of a clause.64% similarUnverifiedMechanically splitting documents by a fixed token count causes headings and the paragraphs they belong to to be placed in different chunks, stripping the paragraph of key disambiguating information such as which package it belongs to63% similarUnverifiedThere are 14 notes that were never opened across 393 sessions, but two of them are in the always-loaded index, so deleting them would lose knowledge.63% similarUnverifiedA read returning an empty value is insufficient to determine that an external write operation failed, because the read itself may be subject to delay or inconsistency.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1048026API
curl https://kongchang.com/api/v1/knowledge/claims/1048026MCP
get_claim(id=1048026)