Unverified50% confidenceFactExact time
The MoE architecture divides the model into multiple expert sub-networks, activating only a small portion during each inference, reducing computational costs while maintaining knowledge capacity; Google's Gemini and Mistral's Mixtral series both adopt this architecture.
1
Sources
50%
Confidence
Long-term
Relevance
9/10/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedMoE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80% similarVerifiedThe 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.76% similarUnverifiedOpenAI、Google 等公司都在探索混合专家（MoE）架构，通过推理时只激活部分参数来降低计算开销74% 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.69% similarUnverifiedModels such as Meta's Llama 3 and Google's Gemini have begun using large amounts of AI-generated data to train the next generation of models.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/887375API
curl https://kongchang.com/api/v1/knowledge/claims/887375MCP
get_claim(id=887375)