<?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>Grokking on AmeyArc</title><link>https://amey-thakur.github.io/tags/grokking/</link><description>Recent content in Grokking on AmeyArc</description><generator>Hugo -- 0.152.2</generator><language>en-us</language><lastBuildDate>Wed, 12 Aug 2026 19:59:59 -0400</lastBuildDate><atom:link href="https://amey-thakur.github.io/tags/grokking/index.xml" rel="self" type="application/rss+xml"/><item><title>Can a Neural Network Learn Exact Arithmetic?</title><link>https://amey-thakur.github.io/posts/2026-08-12-can-a-neural-network-learn-exact-arithmetic/</link><pubDate>Wed, 12 Aug 2026 19:59:59 -0400</pubDate><guid>https://amey-thakur.github.io/posts/2026-08-12-can-a-neural-network-learn-exact-arithmetic/</guid><description>A calculator multiplies two thousand-digit numbers and takes the remainder without thinking about it. Ask a neural network to do the same thing and it becomes an open research question, one that Terence Tao helped set as a competition. The answer has to be exact, because a remainder that is off by one is simply wrong. This is what I submitted to the SAIR Foundation&amp;rsquo;s Modular Arithmetic Challenge, why the obvious approach cannot work, and the one idea that made the difference.</description></item></channel></rss>