English
← Back to AI Technology

Inference and deployment

KV Cache

A cache of attention keys and values from earlier tokens during autoregressive generation. It avoids recomputing the full prefix but often becomes a major memory consumer.

Reference period
2017–present
Last reviewed

Key terms

  • keys
  • values
  • prefill
  • decode

Connected across the project

Primary sources

  1. Transformers cache explanation

This is a curated technical map, not a claim of comprehensive coverage.