Unverified50% confidenceFactExact time
Opposite items that are reverse-scored within the same dimension are highly similar semantically (semantic similarity .58), almost on par with cross-dimension item pairs (.57), yet their responses are opposite.
1
Sources
50%
Confidence
Long-term
Relevance
10/28/2026
First Seen
Sources
Related Entities
Related Claims
Unverified分形是一类具有自相似性、非整数分形维数的复杂几何对象，其分形维数（Hausdorff维数）可以取非整数值60% similarUnverifiedThe embedding spaces of different models are highly consistent in terms of relative geometric relationships (relative distances, angles, and topological structures), and this consistency is referred to as the universal geometry of embeddings.60% similarUnverifiedThe vast majority of embedding models are anisotropic—their vectors are not uniformly distributed across a hypersphere but are instead clustered within a narrow cone, causing the cosine similarity of unrelated texts to be inflated.58% similarUnverified难抵极点从数学本质上等价于求解区域内切圆的圆心，在计算几何中被称为Chebyshev中心问题57% similarUnverifiedK-means assumes clusters are convex and isotropic, performs poorly on non-spherical distributions, and is extremely sensitive to outliers.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1105886API
curl https://kongchang.com/api/v1/knowledge/claims/1105886MCP
get_claim(id=1105886)