Unverified50% confidenceFactExact time
MoE models such as DeepSeek and Mixtral adopt the "large total, small activation" naming convention.
1
Sources
50%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedDeepSeek's key innovation lies in its fine-grained expert splitting strategy, which breaks down the larger experts in traditional MoE into a greater number of smaller units.75% similarUnverifiedLarge model development relies on the comprehension capabilities of underlying large models, including models such as DeepSeek, Kimi, Doubao, and GPT.70% similarUnverifiedBoth DeepSeek and Mistral's Mixtral series adopt the MoE architecture.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1032614API
curl https://kongchang.com/api/v1/knowledge/claims/1032614MCP
get_claim(id=1032614)