Unverified50% confidenceFactExact time
Contextual embedding introduces surrounding text or full-document information as conditional input when generating vectors, causing the same passage to produce different vector representations in different articles.
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedRunning K-Means directly on high-dimensional text embeddings may cause clustering results to be dominated by a small number of vectors with extremely large magnitudes due to differences in document embedding magnitudes.71% similarUnverifiedVerb-final structure allows writers to stack large amounts of qualifying and modifying information before the verb, helping to accommodate complex logical relationships.69% similarUnverified提示词注入分为直接注入和间接注入两种形式，因语言模型无法区分数据与指令，凡进入上下文窗口的文本都可能被执行68% similarUnverified仅仅改变输入文本的措辞和结构，就能让同一个模型的性能产生数十个百分点的差异68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/975239API
curl https://kongchang.com/api/v1/knowledge/claims/975239MCP
get_claim(id=975239)