<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hojun Lee — Writing</title><description>Research notes and paper reviews on CS, ML and AI.</description><link>https://hojunlee.me/writing/</link><language>en</language><atom:link href="https://hojunlee.me/rss.xml" rel="self" type="application/rss+xml"/><item><title>LLM Context</title><link>https://hojunlee.me/ko/writing/llm-context/</link><guid isPermaLink="true">https://hojunlee.me/ko/writing/llm-context/</guid><description>How a chat context is kept as a list of role-tagged messages, tried on Qwen3-1.7B.</description><pubDate>Sun, 19 Jul 2026 00:00:00 GMT</pubDate><category>llm</category><category>notes</category></item><item><title>LLMs Get Lost In Multi-Turn Conversation</title><link>https://hojunlee.me/ko/writing/llms-get-lost-in-multi-turn-conversation/</link><guid isPermaLink="true">https://hojunlee.me/ko/writing/llms-get-lost-in-multi-turn-conversation/</guid><description>Paper notes: measuring the gap between single-turn evals and real multi-turn use by sharding instructions.</description><pubDate>Sun, 19 Jul 2026 00:00:00 GMT</pubDate><category>paper-review</category><category>llm</category><category>evaluation</category></item></channel></rss>