<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Knowledge Transfer from High-Resource to Low-Resource Programming Languages for Code LLMs · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2308.09895/</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/2308.09895/feed.xml" rel="self" type="application/rss+xml" /><item><title>Knowledge Transfer from High-Resource to Low-Resource Programming Languages for Code LLMs — HN submission</title><link>https://news.ycombinator.com/item?id=45662886</link><guid isPermaLink="false">urn:hackerpapers:submission:2308.09895:45662886</guid><pubDate>Tue, 21 Oct 2025 23:05:24 GMT</pubDate><description>&lt;p&gt;Knowledge Transfer from High-Resource to Low-Resource Languages for Code LLMs (2023)&lt;/p&gt;&lt;p&gt;&lt;/p&gt;&lt;p&gt;1 point; 0 discussion replies on HN. &lt;a href="https://hackerpaper.kaipfister.de/p/2308.09895/"&gt;Paper and source history&lt;/a&gt;. A submission is not an endorsement.&lt;/p&gt;</description><category>cs.PL</category></item><item><title>Knowledge Transfer from High-Resource to Low-Resource Programming Languages for Code LLMs — mentioned by groby_b</title><link>https://news.ycombinator.com/item?id=43418766</link><guid isPermaLink="false">urn:hackerpapers:mention:2308.09895:43418766</guid><pubDate>Thu, 20 Mar 2025 00:33:33 GMT</pubDate><description>&lt;p&gt;&amp;gt; the guys I get are pretty good and actually learn. The model doesn&amp;#x27;t. Core issue. LLMs never ever leave their base level unless you actively modify the prompt. I suppose you _could_ use finetuning to whip it into a useful shape, but that&amp;#x27;s a lot of work. (https://arxiv.org/pdf/2308.09895 is a good read) But the flip side of that core issue is that if the base level is high, they&amp;#x27;re good. Which means for Python &amp;amp; JS, they&amp;#x27;re pretty darn good. Making pandas garbage work? Just the task for an LLM. But yeah, R &amp;amp; nginx is not a major part of their original training data, and so they&amp;#x27;re stuck at…&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2308.09895/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.PL</category></item></channel></rss>
