<?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>Interpretability on TechLLM</title><link>https://techllm.github.io/tags/interpretability/</link><description>Recent content in Interpretability on TechLLM</description><generator>Hugo</generator><language>ko</language><lastBuildDate>Thu, 30 Apr 2026 16:35:00 +0900</lastBuildDate><atom:link href="https://techllm.github.io/tags/interpretability/index.xml" rel="self" type="application/rss+xml"/><item><title>Anthropic이 Claude의 생각을 추적해 발견한 것들</title><link>https://techllm.github.io/posts/anthropic-tracing-thoughts-language-model/</link><pubDate>Thu, 30 Apr 2026 16:35:00 +0900</pubDate><guid>https://techllm.github.io/posts/anthropic-tracing-thoughts-language-model/</guid><description>Anthropic의 회로 추적 연구를 바탕으로 Claude가 언어를 넘나들고, 미리 계획하고, 때로는 그럴듯한 가짜 추론을 만드는 방식을 핵심만 정리한다.</description></item></channel></rss>