<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>GLU on Nalar</title>
    <link>https://nalar.dev/tags/glu/</link>
    <description>Recent content in GLU on Nalar</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Mon, 01 Jan 0001 00:00:00 +0000</lastBuildDate>
    <atom:link href="https://nalar.dev/tags/glu/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Use Gated Feed-Forward Blocks in Transformers</title>
      <link>https://nalar.dev/use-gated-feed-forward-blocks-in-transformers/</link>
      <pubDate>Fri, 11 Sep 2026 00:00:00 +0700</pubDate>
      <guid>https://nalar.dev/use-gated-feed-forward-blocks-in-transformers/</guid>
      <description>&lt;h1 id=&#34;use-gated-feed-forward-blocks-in-transformers&#34;&gt;Use Gated Feed-Forward Blocks in Transformers&lt;/h1&gt;&#xA;&lt;p&gt;A transformer block does more than attention. After attention mixes information across token positions, a feed-forward block transforms each token representation independently. That second stage often accounts for a large share of the model&amp;rsquo;s parameters and computation.&lt;/p&gt;&#xA;&lt;p&gt;A gated feed-forward block changes this stage by adding a second projection that controls how much of another projected signal passes through. GLU, GEGLU, and SwiGLU are common names for members of this family.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
