Unverified50% confidenceFactExact time
GQA由Google在2023年提出，将多个查询头共享同一组Key-Value头，已被LLaMA 3、Mistral等主流模型采用
1
Sources
50%
Confidence
Long-term
Relevance
7/6/2026
First Seen
Sources
Needle: Fine-Tuning a 26M Ultralight Tool-Calling Model Locally — Full Walkthrough
bilibili薛定猫AI7/3/2026
Related Claims
UnverifiedGQA由Google在2023年提出，将多个查询头共享同一组Key-Value头，可降低KV Cache内存消耗50%-75%82% similarUnverifiedGQA (Grouped Query Attention) is an efficient variant of standard multi-head attention, proposed by Google in 2023, and has been adopted by mainstream models such as LLaMA 3 and Mistral.81% similarUnverifiedGoogle released the Gemma 4 series of open-source models, comprising four models of different parameter scales.71% similarUnverifiedGQA was first proposed by Google during the training of PaLM 2 and subsequently adopted by open-source models such as Llama 2/3 and Mistral.69% similarVerified思维链（Chain-of-Thought）提示技术由Google Research于2022年提出69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/112995API
curl https://kongchang.com/api/v1/knowledge/claims/112995MCP
get_claim(id=112995)