<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Memory Transformer · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2006.11527/</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/2006.11527/feed.xml" rel="self" type="application/rss+xml" /><item><title>Memory Transformer — mentioned by thomasahle</title><link>https://news.ycombinator.com/item?id=36166227</link><guid isPermaLink="false">urn:hackerpapers:mention:2006.11527:36166227</guid><pubDate>Fri, 02 Jun 2023 15:11:58 GMT</pubDate><description>&lt;p&gt;Have you tried something like Memory Transformers https://arxiv.org/abs/2006.11527 where you move the k/v pairs that don&amp;#x27;t fit in the context window to a vector db? Seems like a more general approach, but I have tested then against each other.&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2006.11527/"&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>
