<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Thresholds on Programmer.ie: Modern AI programming</title>
    <link>http://programmer.ie/tags/thresholds/</link>
    <description>Recent content in Thresholds on Programmer.ie: Modern AI programming</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Mon, 07 Sep 2026 12:10:00 +0000</lastBuildDate>
    <atom:link href="http://programmer.ie/tags/thresholds/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Calibration</title>
      <link>http://programmer.ie/books/embeddings-from-first-principles/14-chapter/</link>
      <pubDate>Mon, 07 Sep 2026 12:10:00 +0000</pubDate>
      <guid>http://programmer.ie/books/embeddings-from-first-principles/14-chapter/</guid>
      <description>&lt;p&gt;&lt;em&gt;Part IV — Measuring the Representation&lt;/em&gt;&lt;/p&gt;&#xA;&lt;h2 id=&#34;what-is-081&#34;&gt;What is 0.81?&lt;/h2&gt;&#xA;&lt;p&gt;A pipeline decides two documents are &amp;ldquo;duplicates&amp;rdquo; if their cosine exceeds 0.8. Someone picked 0.8 because it looked reasonable. Then:&lt;/p&gt;&#xA;&lt;div class=&#34;highlight&#34;&gt;&lt;pre tabindex=&#34;0&#34; style=&#34;color:#f8f8f2;background-color:#272822;-moz-tab-size:4;-o-tab-size:4;tab-size:4;&#34;&gt;&lt;code class=&#34;language-text&#34; data-lang=&#34;text&#34;&gt;&lt;span style=&#34;display:flex;&#34;&gt;&lt;span&gt;distribution of cosine for KNOWN duplicate pairs:      mean 0.79, sd 0.09&#xA;&lt;/span&gt;&lt;/span&gt;&lt;span style=&#34;display:flex;&#34;&gt;&lt;span&gt;distribution of cosine for KNOWN non-duplicate pairs:  mean 0.71, sd 0.11&#xA;&lt;/span&gt;&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;&lt;/div&gt;&lt;p&gt;The threshold sits inside both distributions. A true duplicate at 0.78 is rejected; an unrelated pair at 0.82 is accepted. The number 0.81 does not carry its own meaning — it is a point on two overlapping bell curves, and without those curves it is noise.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
