<?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>PromptEngineering on SiBlog</title><link>https://sinimite.work/tags/promptengineering/</link><description>Recent content in PromptEngineering on SiBlog</description><image><title>SiBlog</title><url>https://sinimite.work/images/og-default.svg?v=20260525-210321</url><link>https://sinimite.work/images/og-default.svg?v=20260525-210321</link></image><generator>Hugo -- 0.156.0</generator><language>zh-cn</language><lastBuildDate>Sun, 26 Jul 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://sinimite.work/tags/promptengineering/rss.xml" rel="self" type="application/rss+xml"/><item><title>生产级 AI Agent 的上下文压缩：从消息裁剪、工具结果清理到 Hermes Agent 的分层 Compaction</title><link>https://sinimite.work/posts/ai-agent-context-compression-best-practices-2026/</link><pubDate>Sun, 26 Jul 2026 00:00:00 +0000</pubDate><guid>https://sinimite.work/posts/ai-agent-context-compression-best-practices-2026/</guid><description>一份面向生产级 AI Agent 的上下文压缩指南，覆盖上下文退化、工具结果管理、Artifact、任务检查点、供应商原生 Compaction、Hermes Agent 实现与工程评估。</description></item><item><title>腾讯 AI Leader 姚顺雨 在这个PodCast中聊了什么？</title><link>https://sinimite.work/posts/points-of-the-podcast-language-agents-from-reasoning-to-acting/</link><pubDate>Sun, 24 May 2026 16:05:10 +0900</pubDate><guid>https://sinimite.work/posts/points-of-the-podcast-language-agents-from-reasoning-to-acting/</guid><description>一篇关于 Language Agents: From Reasoning to Acting podcast 的观后感，从 ReAct、Reflexion、Tree of Thoughts、memory、benchmark 和 ACI 出发，解释 LLM agent 为什么是模型、工具、记忆、环境、评估和 UX 共同构成的工程系统。</description></item><item><title>LLM 的“心口不一”: 理解 LLM 行为的一把万能钥匙</title><link>https://sinimite.work/posts/llm-from-hidden-state-to-token-output/</link><pubDate>Sat, 02 May 2026 00:00:00 +0000</pubDate><guid>https://sinimite.work/posts/llm-from-hidden-state-to-token-output/</guid><description>为什么 Chain of Thought 有效，为什么 prompt 不是越长越好，为什么模型会幻觉？这些看似无关的工程现象，背后其实是同一个核心机制在不同侧面的体现。本文用一把“万能钥匙”打开 LLM 工程的五扇门。</description></item><item><title>LLM 中的熵：从训练到推理到产品的统一语言</title><link>https://sinimite.work/posts/entropy-in-llm/</link><pubDate>Fri, 27 Mar 2026 00:00:00 +0000</pubDate><guid>https://sinimite.work/posts/entropy-in-llm/</guid><description>熵（Entropy）贯穿 LLM 的整个生命周期——训练用交叉熵做 loss，推理用 temperature 控制输出熵，评估用 perplexity 衡量质量，产品层面用 entropy 信号检测幻觉。本文从一个转型 AI 工程师的视角，梳理熵在 LLM 中的完整图谱。</description></item><item><title>Prompt Engineering 完全最佳实践指南</title><link>https://sinimite.work/posts/prompt-engineering-best-practices-guide/</link><pubDate>Wed, 11 Mar 2026 11:00:00 +0900</pubDate><guid>https://sinimite.work/posts/prompt-engineering-best-practices-guide/</guid><description>系统梳理 2025-2026 年 Prompt Engineering 与 Context Engineering 的生产级最佳实践，涵盖 Claude、GPT、长上下文、Prompt Chaining、Tool Use 与 Eval 工作流。</description></item><item><title>LLM Chain-of-Thought (CoT) 完全指南</title><link>https://sinimite.work/posts/llm-chain-of-thought-cot/</link><pubDate>Wed, 04 Mar 2026 21:51:00 +0900</pubDate><guid>https://sinimite.work/posts/llm-chain-of-thought-cot/</guid><description>理解什么是LLM Chain-of-Thought (CoT)，以及如何进行 prompt engineering触发 LLM的Chain-of-Thought (CoT)</description></item><item><title>Prompt Engineering从原理到实战</title><link>https://sinimite.work/posts/prompt-engineering-from-concept-to-implementation/</link><pubDate>Tue, 03 Mar 2026 19:50:00 +0900</pubDate><guid>https://sinimite.work/posts/prompt-engineering-from-concept-to-implementation/</guid><description>提示词工程？也许没有你想的那么简单。</description></item><item><title>Prompt 注入</title><link>https://sinimite.work/posts/understanding-prompt-injection/</link><pubDate>Tue, 03 Mar 2026 19:50:00 +0900</pubDate><guid>https://sinimite.work/posts/understanding-prompt-injection/</guid><description>了解什么事 LLM的 Prompt注入，以及了解一些最近本的防御措施</description></item></channel></rss>