5 comments

  • skybrian9 minutes ago
    Previously posted to Anthropic&#x27;s blog in October. (Maybe not the same version?)<p><a href="https:&#x2F;&#x2F;www.anthropic.com&#x2F;research&#x2F;introspection" rel="nofollow">https:&#x2F;&#x2F;www.anthropic.com&#x2F;research&#x2F;introspection</a>
  • Retro_Dev1 hour ago
    &gt; Overall, our results indicate that current language models possess some functional introspective awareness of their own internal states. We stress that <i>in today’s models, this capacity is highly unreliable and context-dependent;</i> however, it may continue to develop with further improvements to model capabilities.
    • Escapade516045 minutes ago
      If it&#x27;s context dependent is it really introspection?
  • pyaamb24 minutes ago
    In thinking of directions where LLM&#x27;s could develop from here, I cant help but think that a models ability to self introspect would immensely improve their utility. The R&amp;D on how to achieve that is beyond me though. How do you train someone how to introspect? Also would it require a continuous learning architecture that doesn&#x27;t separate training and inference?
  • whythismatters2 hours ago
    &gt;Submitted on 5 Jan 2026
  • pevansgreenwood17 minutes ago
    [dead]