Concept因果编码架构 / Causal Encoding Decoding
CED Architecture
An asymmetric neural network architecture proposed by DeepSeek in which different numbers of parameters are activated on the input and output sides. This design reduces computational waste when processing large inputs and is particularly suited for AI Agent scenarios where inputs significantly outnumber outputs.
Core Facts
Timeline (last 90 days)
Sep 11
V4.1 Flash adopts a new CED (Causal Encoding) architecture, whose core idea is an asymmetry in computation between input and output.
Verified70%
Sep 11
In the CED architecture, only 8B parameters are activated on the input side, while 16B parameters are activated on the output side.
Unverified50%