Model techniques map
Techniquesinference & servingKV cache management

specific method · filed under inference & serving

Least-recently-used eviction

Evicts cached entries according to least-recently-used order.

Also called LRU eviction policy.

source
1
model
1
lab adopt it
1
strongest
used

How sources treat it

One count per evidence span, weakest treatment to strongest.

used 1

Documented in

Evidence

1 span quoted from the sources, strongest treatment first.

global KV and SWA KV are managed independently, governed by an LRU eviction policy

usedunclearin DeepSeek-V4.1-FlashDeepSeek

Filed alongside

Other methods under inference & serving :: KV cache management.