<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Compressing LLMs: The Truth is Rarely Pure and Never Simple · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2310.01382/</link><description>Collected HN comments and submissions linking this paper.</description><language>en</language><lastBuildDate>Mon, 02 Nov 2026 11:16:22 GMT</lastBuildDate><atom:link href="https://hackerpaper.kaipfister.de/p/2310.01382/feed.xml" rel="self" type="application/rss+xml" /><item><title>Compressing LLMs: The Truth is Rarely Pure and Never Simple — mentioned by janwas</title><link>https://news.ycombinator.com/item?id=40817984</link><guid isPermaLink="false">urn:hackerpapers:mention:2310.01382:40817984</guid><pubDate>Fri, 28 Jun 2024 05:22:45 GMT</pubDate><description>&lt;p&gt;Yes, PPL and certain benchmarks do not detect differences from quantization. But recent work gives cause for concern, e.g., https://arxiv.org/pdf/2310.01382, https://arxiv.org/pdf/2405.18137.&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2310.01382/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CL</category></item></channel></rss>
