<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Sparse-Attention on Zhanyl Abdybaeva</title>
    <link>https://zhanyl-tech.github.io/tags/sparse-attention/</link>
    <description>Recent content in Sparse-Attention on Zhanyl Abdybaeva</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Fri, 30 Jan 2026 00:00:00 +0000</lastBuildDate>
    <atom:link href="https://zhanyl-tech.github.io/tags/sparse-attention/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>LServe and SampleAttention: What Sparse Attention Actually Changes in Prefill and Decode</title>
      <link>https://zhanyl-tech.github.io/blog/2026-01-30-lserve-sampleattention-sparse-attention/</link>
      <pubDate>Fri, 30 Jan 2026 00:00:00 +0000</pubDate>
      <guid>https://zhanyl-tech.github.io/blog/2026-01-30-lserve-sampleattention-sparse-attention/</guid>
      <description>Structured sparsity aligned with GPU attention kernels: LServe’s unified serving stack for prefill and decode, and SampleAttention’s empirical patterns plus CRA as a runtime quality floor.</description>
    </item>
  </channel>
</rss>
