AI Inference & Serving

Multi-Query Attention

An attention variant that retains multiple query heads while sharing one key head and one value head across them. The sharing changes the stored key/value representation, rather than reducing a conversation to one query. It is an architectural choice, not a switch guaranteed to preserve any existing model unchanged.

Reviewed

Sources

Member lesson

The definition and sources are public. The complete practical lesson is for members.

Compare membership plans ยท Already a member? Sign in

Explore all dictionary definitions