<?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>Evals on Local First AI</title><link>https://localfirstai.eu/tags/evals/</link><description>Recent content in Evals on Local First AI</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Sat, 06 Jun 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://localfirstai.eu/tags/evals/index.xml" rel="self" type="application/rss+xml"/><item><title>The Adversarial Watcher: When a Local Model Audits Its Own Project</title><link>https://localfirstai.eu/posts/2026-06-06-adversarial-watcher/</link><pubDate>Sat, 06 Jun 2026 00:00:00 +0000</pubDate><guid>https://localfirstai.eu/posts/2026-06-06-adversarial-watcher/</guid><description>Documentation drifts from code. A staged local LLM pipeline — five steps, each under 4K tokens — systematically catches the gap. First results, false positive anatomy, and an open question about what comes next.</description></item></channel></rss>