Open Models Almanac
Methodsinference & servingKV cache management

single mention · inference & serving

Prefix key-value caching during action inference

Caches attention keys and values for the prefix to make action inference more efficient.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See KV cache management for the methods filed beside it.

Filed under inference & serving › KV cache management