{"id":146353,"date":"2026-07-08T09:14:00","date_gmt":"2026-07-08T13:14:00","guid":{"rendered":"https:\/\/kendallharmon.net\/?p=146353"},"modified":"2026-07-08T18:34:02","modified_gmt":"2026-07-08T22:34:02","slug":"anthropic-a-global-workspace-in-language-models","status":"publish","type":"post","link":"https:\/\/kendallharmon.net\/?p=146353","title":{"rendered":"(Anthropic) A global workspace in language models"},"content":{"rendered":"\n

As you read this sentence, circuits in your brain are adjusting your posture, controlling your breathing, and transforming lines and curves on the screen into recognizable words. Most of this processing is invisible to you. But some of what takes place in your brain you do<\/em> have access to\u2014an image that pops into your head, or a deliberate plan you make about where to go shopping. Neuroscientists and philosophers sometimes refer to the latter type of brain activity as \u201cconsciously accessible,\u201d to distinguish it from all the other processing that goes on unconsciously. This activity has special properties: we can describe it, control it, and use it for deliberate reasoning, in contrast to all the automatic processing that goes on without our awareness.<\/p>\n\n\n\n

In a new paper, we present evidence that a similar distinction has emerged in modern language models like Claude. We find that Claude has developed a small collection of internal neural patterns that, compared to all its other internal processing, play a special role.<\/p>\n\n\n\n

We call the collection of these patterns the J-space<\/em>\u2014named after the technique we used to find them, involving a mathematical concept called the Jacobian. Each J-space pattern is linked to a particular word. But when one of these patterns lights up, it doesn\u2019t mean the model is saying<\/em> that word\u2014just that the word is on its mind. If you’ve heard of language models having a “scratchpad” or \u201cchain of thought\u201d\u2014text they write to themselves while reasoning\u2014the J-space is something different. It operates silently, in the model\u2019s internal neural activations, allowing the model to think about a concept without writing it down. Notably, the J-space wasn\u2019t designed or programmed by us, but instead emerged on its own<\/em> during Claude\u2019s training process.<\/p>\n\n\n\n

Read it all<\/a>.<\/p>

Wow! Unexpected results from Anthropic's experiments in exposing the inner thoughts of Claude. They call its unconscious a J-space. The short video they made is good place to start. https:\/\/t.co\/07cr8aJlxr<\/a><\/p>— Kevin Kelly (@kevin2kelly) July 8, 2026<\/a><\/blockquote>