---
cited_by:
  - '[[thoughts/inference/architectures#6-reading-order|README §6]]'
  - '[[thoughts/inference/kimi-k3-state#32-prefix-caching-for-recurrent-state|kimik3 §3.2]]'
  - '[[thoughts/inference/kimi-k3-prs#4-pr-reading-order|kimik3 §4]]'
  - '[[thoughts/inference/kimi-k3#5-sharp-edges-each-fix-shows-an-invariant|kimik3 §5]]'
date: '2026-11-11'
kind: pr
local:
  - file:///Users/aarnphm/workspace/vllm/vllm/v1/core/single_type_kv_cache_manager.py
merged: '2026-09-15'
number: 56794
source:
  - https://github.com/vllm-project/vllm/blob/f547c23ec971fd1c78f0d89d15ac1957ae407471/vllm/v1/core/single_type_kv_cache_manager.py
state: merged
tags:
  - vllm
  - vllm-pr
  - new-architectures
title: '#56794 [Bugfix][Kimi-K3] Keep transient checkpoints out of prefix-cache eviction'
topic: new-architectures
url: https://github.com/vllm-project/vllm/pull/56794
created: '2026-11-11'
modified: '2026-11-11'
published: '2026-11-11'
pageLayout: default
slug: hinterland/papers/maps/prs/56794
permalink: https://aarnphm.xyz/hinterland/papers/maps/prs/56794.md
generator:
  quartz: v4.6.0
  hostedProvider: Cloudflare
  baseUrl: aarnphm.xyz
full: https://aarnphm.xyz/llms-full.txt
---
Keep transient KDA/Mamba prefill checkpoints out of the ordinary prefix-cache eviction set when sparse retention is explicitly active. · full dump: [[thoughts/inference/vllm-prs/56794|#56794]]

