Unverified50% confidenceFactExact time
DeepSeek V2, released in May 2024, adopted the self-developed MLA attention mechanism, cutting inference costs to one-tenth of the original.
1
Sources
50%
Confidence
Long-term
Relevance
10/23/2026
First Seen
Sources
Understanding DeepSeek in 3 Minutes: The Rise from a Quant Hedge Fund to an Open-Source AI Leader
bilibiliGerrardYX10/23/2026
Related Entities
Related Claims
UnverifiedDeepSeek-V2引入Multi-Head Latent Attention（MLA）机制和DeepSeekMoE架构，其每百万token推理成本约为同期GPT-4级别模型的百分之一78% similarUnverified2022年DeepMind发布了TALM，Meta发布了Toolformer，探索让模型自主决定何时调用外部API，但依赖微调70% similarUnverifiedDeepSeek在2024年初以DeepSeek-V2的超低推理成本定价引发行业震动，DeepSeek-R1在推理能力上取得突破性表现68% similarVerifiedDeepSeek V4 adopts a proprietary hierarchical compressed attention strategy, featuring two alternating attention modes: CSA (Compressed Sink Attention) and HCA (Heavy Compressed Attention).68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1090497API
curl https://kongchang.com/api/v1/knowledge/claims/1090497MCP
get_claim(id=1090497)