<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>LLMs are Bayesian, In Expectation, Not in Realization · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2507.11768/</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/2507.11768/feed.xml" rel="self" type="application/rss+xml" /><item><title>LLMs are Bayesian, In Expectation, Not in Realization — mentioned by rishabhaiover</title><link>https://news.ycombinator.com/item?id=47308281</link><guid isPermaLink="false">urn:hackerpapers:mention:2507.11768:47308281</guid><pubDate>Mon, 09 Mar 2026 12:38:51 GMT</pubDate><description>&lt;p&gt;Damn, the research moves fast. I was wrong again: https://arxiv.org/abs/2507.11768&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2507.11768/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>stat.ML</category></item><item><title>LLMs are Bayesian, In Expectation, Not in Realization — mentioned by spindump8930</title><link>https://news.ycombinator.com/item?id=45182096</link><guid isPermaLink="false">urn:hackerpapers:mention:2507.11768:45182096</guid><pubDate>Tue, 09 Sep 2025 14:04:12 GMT</pubDate><description>&lt;p&gt;This topic is interesting, but the repo and paper have a lot of inconsistencies that make me think this work is hiding behind lots of dense notation and language. For one, the repo states: &amp;gt; This implementation follows the framework from the paper “Compression Failure in LLMs: Bayesian in Expectation, Not in Realization” (NeurIPS 2024 preprint) and related EDFL/ISR/B2T methodology. There doesn&amp;#x27;t seem to be a paper by that title, preprint or actual neurips publication. There is https://arxiv.org/abs/2507.11768, with a different title, and contains lots of inconsistencies with regards to the…&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2507.11768/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>stat.ML</category></item><item><title>LLMs are Bayesian, In Expectation, Not in Realization — mentioned by fiduciarytemp</title><link>https://news.ycombinator.com/item?id=45181114</link><guid isPermaLink="false">urn:hackerpapers:mention:2507.11768:45181114</guid><pubDate>Tue, 09 Sep 2025 12:37:03 GMT</pubDate><description>&lt;p&gt;Paper: https://arxiv.org/abs/2507.11768&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2507.11768/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>stat.ML</category></item><item><title>LLMs are Bayesian, In Expectation, Not in Realization — HN submission</title><link>https://news.ycombinator.com/item?id=44668806</link><guid isPermaLink="false">urn:hackerpapers:submission:2507.11768:44668806</guid><pubDate>Thu, 24 Jul 2025 09:39:43 GMT</pubDate><description>&lt;p&gt;LLMs are Bayesian, in Expectation, not in Realization&lt;/p&gt;&lt;p&gt;&lt;/p&gt;&lt;p&gt;2 points; 0 discussion replies on HN. &lt;a href="https://hackerpaper.kaipfister.de/p/2507.11768/"&gt;Paper and source history&lt;/a&gt;. A submission is not an endorsement.&lt;/p&gt;</description><category>stat.ML</category></item></channel></rss>
