<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Season-4 on GPUburnout | Jun Park</title><link>https://gpuburnout.com/tags/season-4/</link><description>Recent content in Season-4 on GPUburnout | Jun Park</description><image><title>GPUburnout | Jun Park</title><url>https://gpuburnout.com/images/og-default.png</url><link>https://gpuburnout.com/images/og-default.png</link></image><generator>Hugo -- 0.155.2</generator><language>en-us</language><lastBuildDate>Mon, 30 Mar 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://gpuburnout.com/tags/season-4/index.xml" rel="self" type="application/rss+xml"/><item><title>Verbatim: The Proof Is in the Output</title><link>https://gpuburnout.com/posts/s4-ch4-verbatim-the-proof-is-in-the-output/</link><pubDate>Mon, 30 Mar 2026 00:00:00 +0000</pubDate><guid>https://gpuburnout.com/posts/s4-ch4-verbatim-the-proof-is-in-the-output/</guid><description>Side-by-side verbatim outputs from the 1B and 2B models. Same prompts, same settings. The difference is everything.</description></item><item><title>7 Out of 8 - How DPO Finally Worked</title><link>https://gpuburnout.com/posts/s4-ch3-dpo-finally-worked/</link><pubDate>Sun, 29 Mar 2026 00:00:00 +0000</pubDate><guid>https://gpuburnout.com/posts/s4-ch3-dpo-finally-worked/</guid><description>In which the same technique that failed nine times on the 1B succeeds on the 2B - and the reason why is the whole point of Season 4.</description></item><item><title>1.92B Parameters, 38.4B Tokens, Zero Garbage</title><link>https://gpuburnout.com/posts/s4-ch2-192b-38b-tokens-zero-garbage/</link><pubDate>Sat, 28 Mar 2026 00:00:00 +0000</pubDate><guid>https://gpuburnout.com/posts/s4-ch2-192b-38b-tokens-zero-garbage/</guid><description>In which a larger, cleaner model is built from the ruins of the last one - and the hypothesis from Season 3 starts looking right.</description></item><item><title>RIP GPUburnout-1B. Cause of Death: Its Own Training Data.</title><link>https://gpuburnout.com/posts/s4-ch1-rip-gpuburnout-1b/</link><pubDate>Sun, 22 Mar 2026 00:00:00 +0000</pubDate><guid>https://gpuburnout.com/posts/s4-ch1-rip-gpuburnout-1b/</guid><description>In which the 1B model is officially retired, the autopsy is filed, and a better model begins.</description></item></channel></rss>