[KongchangAI]
Unverified50% confidenceFactExact time

MLA 将所有头的 Key 和 Value 联合压缩为低维潜在向量缓存,本质上是对 KV Cache 做低秩分解

1
Sources
50%
Confidence
Long-term
Relevance
7/14/2026
First Seen

Sources

Related Claims

Cite This Claim

Stable URI
https://kongchang.com/claim/503163
API
curl https://kongchang.com/api/v1/knowledge/claims/503163
MCP
get_claim(id=503163)