---
title: "Multi-query attention mathematical notation"
description: "Multi-query attention is a recurring research-paper notation family. Uses many query heads but shares one key head and one value head across all of them…"
canonical_url: "https://fanout.sh/labs/math-decoder/symbol/multi-query-attention"
md_url: "https://fanout.sh/labs/math-decoder/symbol/multi-query-attention.md"
last_updated: "2026-08-09"
access: "public"
---

# Multi-query attention mathematical notation

Multi-query attention is a recurring research-paper notation family. Uses many query heads but shares one key head and one value head across all of them…

## Public overview

Multi-query attention is a recurring research-paper notation family. Uses many query heads but shares one key head and one value head across all of them to reduce cache size and memory traffic.

Multi-query attention: Uses many query heads but shares one key head and one value head across all of them to reduce cache size and memory traffic. Example: Reuse the same cached keys and values for every query head.

---
This representation contains public Fanout content only. Protected Pro lessons, account data, billing, checkout, and pricing are not included.

Browse the public content map: https://fanout.sh/sitemap.md