RECORD / 7E8A33 / REVIEWED

Cache-to-Cache

Cache-to-Cache learns projections that fuse a source model's KV cache into a receiver, letting paired language models exchange latent state without generating intermediate text.

State: available
Layer: infrastructure
Reviewed: 2026-10-17

DECISION / USE

Why it matters

Use it to research low-latency cooperation between models when you control both runtimes and can train a cache projector.

CAUTION / LIMITS

What to check first

It requires internal cache access and pair-specific training or compatible checkpoints; reported quality and latency gains are experimental and model-dependent.

REL / KNOWLEDGE

1 related record

01KV Cache Transform Coding →KV Cache Transform Coding applies learned decorrelation, adaptive quantization, and entropy coding to reduce storage and transfer costs for reusable language-model KV caches.research paper