<?xml version='1.0' encoding='utf-8'?>
<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0">
  <channel>
    <title>Minseok Seo · Notes on research</title>
    <link>https://seominseok0429.github.io/blog/</link>
    <description>Minseok Seo’s research notebook: AI safety experiments, research papers, and things tried, shared openly in Korean, English, and Chinese.</description>
    <language>ko</language>
    <atom:link href="https://seominseok0429.github.io/blog/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>OpenAI Astra와 Anthropic Fable이 ‘망상장애에 걸린다면’ 어떻게 행동할까?</title>
      <link>https://seominseok0429.github.io/blog/posts/when-ai-believes-a-false-reality/</link>
      <description>서민석의 AI 안전성 실험: 시각적으로 변환된 장면이 OpenAI Astra와 Anthropic Fable의 판단에 미치는 영향. 원본·변환 이미지, 시뮬레이터 예시와 한계를 공유합니다.</description>
      <category>Things I tried</category>
      <guid isPermaLink="true">https://seominseok0429.github.io/blog/posts/when-ai-believes-a-false-reality/</guid>
    </item>
  </channel>
</rss>