<?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>DeepSeek on heihie</title><link>https://heihei0299.github.io/tags/deepseek/</link><description>Recent content in DeepSeek on heihie</description><generator>Hugo</generator><language>zh-CN</language><lastBuildDate>Mon, 27 Jul 2026 00:00:00 +0800</lastBuildDate><atom:link href="https://heihei0299.github.io/tags/deepseek/index.xml" rel="self" type="application/rss+xml"/><item><title>DeepSeek 长上下文对话的幻觉与使用规范</title><link>https://heihei0299.github.io/posts/ai-issues/deepseek%E4%BD%BF%E7%94%A8%E6%B1%87%E6%80%BB/</link><pubDate>Mon, 27 Jul 2026 00:00:00 +0800</pubDate><guid>https://heihei0299.github.io/posts/ai-issues/deepseek%E4%BD%BF%E7%94%A8%E6%B1%87%E6%80%BB/</guid><description>总结 DeepSeek 在长上下文对话中的三大问题（上下文超限后幻觉加重、上下文混乱、对话结束后擅自写代码）与两条使用规范：避免许愿式编程、提高交流信息密度。</description></item></channel></rss>