Unverified50% confidenceBenchmarkExact time
CLIP ViT-L/14的内在维度约为70-120维，远低于其1024维的标称特征空间
1
Sources
50%
Confidence
Long-term
Relevance
7/9/2026
First Seen
Sources
Block-Sparse Featurizers: Opening the Black Box of Vision Model Neural Geometry
hackernewshackernews7/7/2026
Cite This Claim
Stable URI
https://kongchang.com/claim/349282API
curl https://kongchang.com/api/v1/knowledge/claims/349282MCP
get_claim(id=349282)