<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>大模型对齐 on AI 早报</title><link>https://ai-news.example.com/categories/%E5%A4%A7%E6%A8%A1%E5%9E%8B%E5%AF%B9%E9%BD%90/</link><description>Recent content in 大模型对齐 on AI 早报</description><generator>Hugo</generator><language>zh-CN</language><lastBuildDate>Mon, 20 Jul 2026 23:11:39 +0800</lastBuildDate><atom:link href="https://ai-news.example.com/categories/%E5%A4%A7%E6%A8%A1%E5%9E%8B%E5%AF%B9%E9%BD%90/feed.xml" rel="self" type="application/rss+xml"/><item><title>[EN] GPT-Red：OpenAI 用自我对抗解锁 AI 鲁棒性自增强</title><link>https://ai-news.example.com/articles/en-gpt-redopenai-ai/</link><pubDate>Mon, 20 Jul 2026 23:11:39 +0800</pubDate><guid>https://ai-news.example.com/articles/en-gpt-redopenai-ai/</guid><description>OpenAI 发布了 GPT-Red，一个基于自博弈的自动化红队系统，旨在通过自我对抗训练提升 AI 的安全性、对齐度和对提示注入攻击的鲁棒性。</description></item></channel></rss>