Unverified50% confidenceFactExact time
MoE架构由 Google Brain 于 2017 年提出，GPT-4、Mixtral 8x7B 等知名模型均采用该架构
1
Sources
50%
Confidence
Long-term
Relevance
7/11/2026
First Seen
Sources
智谱GLM-5.2全量开放，AI监管风暴同步来袭
bilibili黑鸦Heya6/13/2026
Related Claims
VerifiedThe Mixture of Experts (MoE) architecture was scaled up within the Transformer framework by Google Brain in 2017; its core idea is to group model parameters into multiple expert sub-networks, with only a small subset activated during each inference.73% similarVerifiedGoogle systematized the Mixture of Experts architecture design in its 2017 Sparsely-Gated MoE paper.73% similarUnverifiedMoE架构最早由Google在2017年提出，后被Mistral的Mixtral系列和阿里的Qwen系列等广泛采用72% similarUnverifiedMoE架构由多个专家子网络和一个路由器组成，路由器根据输入token动态选择少数几个专家参与计算，DeepSeek V2/V3、Mixtral 8x7B、Google Gemini 1.5均采用这一架构72% similarUnverifiedThe MoE architecture was founded on a Google paper from 2017; it dynamically assigns tokens to a small number of expert sub-networks via a routing network, and Qwen 3.5 122B has 122B total parameters but activates only approximately 22B during each inference.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/486910API
curl https://kongchang.com/api/v1/knowledge/claims/486910MCP
get_claim(id=486910)