<?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>Model Composition on Cognaptus</title>
    <link>https://cognaptus.com/tags/model-composition/</link>
    <description>Recent content in Model Composition on Cognaptus</description>
    <generator>Hugo -- 0.145.0</generator>
    <language>en-us</language>
    <lastBuildDate>Sun, 12 Apr 2026 00:00:00 +0000</lastBuildDate>
    <atom:link href="https://cognaptus.com/tags/model-composition/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Dead Weights, Live Signals: When Frozen Models Start Talking</title>
      <link>https://cognaptus.com/blog/2026-04-12-dead-weights-live-signals-when-frozen-models-start-talking/</link>
      <pubDate>Sun, 12 Apr 2026 00:00:00 +0000</pubDate>
      <guid>https://cognaptus.com/blog/2026-04-12-dead-weights-live-signals-when-frozen-models-start-talking/</guid>
      <description>A mechanism-first reading of how frozen language models can be composed through latent-space communication, what the benchmark gains actually support, and where the idea is still fragile.</description>
    </item>
  </channel>
</rss>
