Unverified50% confidenceFactExact time
1.6万亿参数级别的巨型模型通常采用混合专家架构（MoE），每次推理只激活一小部分专家子网络
1
Sources
50%
Confidence
Long-term
Relevance
7/16/2026
First Seen
Sources
Related Claims
VerifiedThe MoE architecture divides a model's internals into multiple expert sub-networks, with a gating network dynamically selecting a small number of experts to participate in computation during each inference, so that the actual number of activated parameters is far smaller than the total parameter count.85% similarUnverifiedMoE架构将模型分割为多个专家子网络，每次推理仅激活少数专家，通讯开销随集群规模呈近似平方级增长82% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/526382API
curl https://kongchang.com/api/v1/knowledge/claims/526382MCP
get_claim(id=526382)