<?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>GRPO on dextermayhewjd</title><link>https://dextermayhewjd.github.io/tags/grpo/</link><description>Recent content in GRPO on dextermayhewjd</description><generator>Hugo</generator><language>zh-cn</language><lastBuildDate>Wed, 07 Oct 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://dextermayhewjd.github.io/tags/grpo/index.xml" rel="self" type="application/rss+xml"/><item><title>Lecture 14 · 从奖励学习到语言模型强化学习</title><link>https://dextermayhewjd.github.io/courses/cs285/lecture-14/</link><pubDate>Wed, 07 Oct 2026 00:00:00 +0000</pubDate><guid>https://dextermayhewjd.github.io/courses/cs285/lecture-14/</guid><description>保留 22 条公式拆解，先展示完整训练循环，再连接 IRL、对抗学习、LLM 策略、GAE/GRPO、参考 KL、偏好与历史状态。</description></item></channel></rss>