Unverified50% confidenceFactExact time
Models extract passages; in a long article, an answer hidden in a later passage is harder to extract than a compact page with a clear heading that gives the answer in the first two sentences.
1
Sources
50%
Confidence
Long-term
Relevance
10/13/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedFor very long documents, a model may have processed only part of the content yet provide an answer that appears complete without any indication.79% similarUnverifiedThe embedding model only sees the sentence pattern of the paragraph and cannot perceive its heading; this is not a flaw of the embedding model, but rather that any fixed-length vector compression discards some information—the problem is that structural clues already provided by the author are being actively discarded75% similarUnverifiedWhen handling long text that exceeds the model's context limit, the text should be split into small chunks, each summarized separately, and then the summaries summarized again.75% similarUnverifiedAs models are optimized for instruction following and content completeness, outputs tend to become more verbose, with a propensity to generate well-structured but uninspired text70% similarUnverified被AI引用的页面倾向于使用问句式标题来匹配用户查询模式，且每章节开头用约25-30词的直接回答实现答案前置，而未被引用的页面首段约为60词的铺垫68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1021129API
curl https://kongchang.com/api/v1/knowledge/claims/1021129MCP
get_claim(id=1021129)