<?xml version="1.0" encoding="utf-8"?><feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en"><generator uri="https://jekyllrb.com/" version="3.10.0">Jekyll</generator><link href="https://nelson960.github.io/Mechanistic-Transparency/feed.xml" rel="self" type="application/atom+xml" /><link href="https://nelson960.github.io/Mechanistic-Transparency/" rel="alternate" type="text/html" hreflang="en" /><updated>2026-04-10T17:34:05+00:00</updated><id>https://nelson960.github.io/Mechanistic-Transparency/feed.xml</id><title type="html">Mechanistic Transparency</title><subtitle>Empirical evidence for staged retrieval motif formation in small transformers trained on controlled in-context retrieval tasks.</subtitle></feed>