<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>How do languages influence each other? Studying cross-lingual data sharing during LM fine-tuning · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2305.13286/</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/2305.13286/feed.xml" rel="self" type="application/rss+xml" /><item><title>How do languages influence each other? Studying cross-lingual data sharing during LM fine-tuning — mentioned by famouswaffles</title><link>https://news.ycombinator.com/item?id=40202951</link><guid isPermaLink="false">urn:hackerpapers:mention:2305.13286:40202951</guid><pubDate>Mon, 29 Apr 2024 19:32:17 GMT</pubDate><description>&lt;p&gt;&amp;gt;I wasn&amp;#x27;t talking about &amp;quot;state of the art LLMs,&amp;quot; I am aware that commercial offerings are much better trained in Spanish. This was a thought experiment based on comments from people testing GPT-3.5 with Swahili. A thought experiment from other people comments on another language. So...No. Fabricating failure modes from their personally constructed ideas about how LLMs work seems to be a frustratingly common occurrence in these kinds of discussions. &amp;gt;Frustratingly, just few months ago I read a paper describing how LLMs excessively rely on English-language representations of ideas, but now I…&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2305.13286/"&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>
