<?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>Calibration on Actually Random</title>
    <link>https://actuallyrandom.com/tags/calibration/</link>
    <description>Recent content in Calibration on Actually Random</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Fri, 10 Jul 2026 21:45:00 +1200</lastBuildDate>
    <atom:link href="https://actuallyrandom.com/tags/calibration/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Can an AI guess what it knows?</title>
      <link>https://actuallyrandom.com/posts/can-an-ai-guess-what-it-knows/</link>
      <pubDate>Fri, 10 Jul 2026 21:45:00 +1200</pubDate>
      <guid>https://actuallyrandom.com/posts/can-an-ai-guess-what-it-knows/</guid>
      <description>I gave a language model 60 New Zealand schools and asked it to guess which ones have Wikipedia articles. It could rank them. It could not put a number on them, and it would have scored better by ignoring everything it knew.</description>
    </item>
  </channel>
</rss>
