<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Bits of Grass: Does GPT already know how to write like Whitman? · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2305.11064/</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/2305.11064/feed.xml" rel="self" type="application/rss+xml" /><item><title>Bits of Grass: Does GPT already know how to write like Whitman? — mentioned by gwern</title><link>https://news.ycombinator.com/item?id=36252179</link><guid isPermaLink="false">urn:hackerpapers:mention:2305.11064:36252179</guid><pubDate>Fri, 09 Jun 2023 01:58:37 GMT</pubDate><description>&lt;p&gt;&amp;gt; However, once they&amp;#x27;ve started writing rhyming poetry it&amp;#x27;s hard to get them to stop rhyming. They seem to have formed a strong association between rhyming and poetry. I&amp;#x27;ve also been unable to get them to obey a specific rhyming scheme like ABBAB. Correct and commonly observed (eg. https://arxiv.org/abs/2305.11064 ). (At least, for GPT models. I don&amp;#x27;t know as much about the Anthropic models as I should, although I understand they do still use a BPE tokenization, unfortunately.) My theory is that it is a surprising interaction of BPEs with RLHF…&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2305.11064/"&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>
