<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title>fiction - Tag - The handbag of Cryptax</title><link>https://cryptax.github.io/tags/fiction/</link><description>fiction - Tag - The handbag of Cryptax</description><generator>Hugo -- gohugo.io</generator><language>en-us</language><managingEditor>crypto.maniak@gmail.com (Cryptax)</managingEditor><webMaster>crypto.maniak@gmail.com (Cryptax)</webMaster><lastBuildDate>Thu, 01 Oct 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://cryptax.github.io/tags/fiction/" rel="self" type="application/rss+xml"/><item><title>Testing JEV against a fictional attack</title><link>https://cryptax.github.io/posts/2026-10-jev/</link><pubDate>Thu, 01 Oct 2026 00:00:00 +0000</pubDate><author>cryptax</author><guid>https://cryptax.github.io/posts/2026-10-jev/</guid><description><![CDATA[I&rsquo;ve finally had a little time to test JEV today.
2026-10-01: Updated with test with Laya
JEV background info If you haven&rsquo;t heard about JEV yet, it&rsquo;s TypeSafe AI&rsquo;s model, which returns typed decisions, with calibrated confidence.
The idea is that, in many cases, when you ask a question to a LLM, you don&rsquo;t care about it&rsquo;s intermediate reasoning and in its final answer, often only 1 or 2 words are really important to you.]]></description></item></channel></rss>