Unverified50% confidenceFactExact time
MoE 架构每次前向传播只激活部分专家子网络，Qwen3.6 35B A3 中约激活 3B 参数
1
Sources
50%
Confidence
Long-term
Relevance
7/22/2026
First Seen
Sources
Related Claims
VerifiedQwen3.6 35B-A3B是混合专家架构（MoE），总参数量35B，每次推理实际激活参数量约3B75% similarUnverified千问3.6的35B MOE模型每次只激活其中3B的参数72% similarPartially VerifiedBoth Qwen 3 Max and DeepSeek-V3 adopt the MoE architecture with total parameter counts reaching hundreds of billions, but the parameters actually activated during inference are only a small fraction of the total.70% similarUnverifiedQwen3 中 MoE 每次激活 K=2 个专家68% similarUnverifiedQwen3 is available on Ollama in multiple sizes including 0.6B, 1.7B, 4B, and 8B67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/585474API
curl https://kongchang.com/api/v1/knowledge/claims/585474MCP
get_claim(id=585474)