<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Language Versatilists vs. Specialists: An Empirical Revisiting on Multilingual Transfer Ability · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2306.06688/</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/2306.06688/feed.xml" rel="self" type="application/rss+xml" /><item><title>Language Versatilists vs. Specialists: An Empirical Revisiting on Multilingual Transfer Ability — mentioned by sanxiyn</title><link>https://news.ycombinator.com/item?id=39175383</link><guid isPermaLink="false">urn:hackerpapers:mention:2306.06688:39175383</guid><pubDate>Mon, 29 Jan 2024 12:12:19 GMT</pubDate><description>&lt;p&gt;https://arxiv.org/abs/2306.06688 studied this. (Check out figure 3.) In short, yes, multilingual transfer does happen. But importantly, not at 7B. Being English-centric is possibly okay for LLaMA 70B, but it really is not okay for LLaMA 7B.&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2306.06688/"&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>
