<?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>Golden-Set on zolty.systems</title><link>https://blog.zolty.systems/tags/golden-set/</link><description>Recent content in Golden-Set on zolty.systems</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Mon, 19 Oct 2026 10:00:00 -0400</lastBuildDate><atom:link href="https://blog.zolty.systems/tags/golden-set/index.xml" rel="self" type="application/rss+xml"/><item><title>Benching a vision pipeline on 17 lots without fooling myself</title><link>https://blog.zolty.systems/posts/2026-10-19-golden-set-vision-bench/</link><pubDate>Mon, 19 Oct 2026 10:00:00 -0400</pubDate><guid>https://blog.zolty.systems/posts/2026-10-19-golden-set-vision-bench/</guid><description>How I measure a lot-reading pipeline with a small golden set, paired McNemar tests and Wilson intervals, and why an accuracy percent on 17 lots is mostly noise.</description></item></channel></rss>