<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>LLMail-Inject: A Dataset from a Realistic Adaptive Prompt Injection Challenge · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2506.09956/</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/2506.09956/feed.xml" rel="self" type="application/rss+xml" /><item><title>LLMail-Inject: A Dataset from a Realistic Adaptive Prompt Injection Challenge — mentioned by mpeg</title><link>https://news.ycombinator.com/item?id=47060410</link><guid isPermaLink="false">urn:hackerpapers:mention:2506.09956:47060410</guid><pubDate>Wed, 18 Feb 2026 12:46:19 GMT</pubDate><description>&lt;p&gt;If you&amp;#x27;re interested in this kind of thing, I took part in a CTF last year organised by Microsoft that was about this exact kind of email injection, with different levels of protection They published the attempts dataset [0] as well as a paper [1] afterwards [0]: https://huggingface.co/datasets/microsoft/llmail-inject-chal... [1]: https://arxiv.org/abs/2506.09956&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2506.09956/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.CR</category></item></channel></rss>
