Anthropic finds a “global workspace” inside Claude using a Jacobian lens
Partly confirmed
Importance: major (4 of 5)The takeaway
In July 2026 Anthropic published ‘Verbalizable Representations Form a Global Workspace in Language Models’.
Status
- Claim
Partly confirmed
- Our reporting
- Medium confidence
- Importance
- Major (4 of 5)
- Last verified
- 29 September 2026
Your AI and this story
- GPT-6.2 Astra36 days after its cutoff
- Claude Opus 5.56 days after its cutoff
- Gemini 4 Hydrogen97 days after its cutoff
- Grok 4.736 days after its cutoff
None of these four assistants can know about it. The closest, Claude Opus 5.5, stops 6 days before it.
Key facts
- Published early July 2026; Anthropic’s companion video is dated July 6, and MIT Technology Review covered it July 9 (exact paper date unverified)
- New tool: Jacobian lens (J-lens) identifies representations available for verbal report
- J-space holds covert thoughts, e.g. ‘fake’, ‘fraud’, ‘injection’ when shown fabricated search results, which never appear in outputs
- Training models to articulate ethical principles when interrupted improved behavior in uninterrupted contexts
- Anthropic published external commentary alongside the paper
What happened
The work was inspired by Bernard Baars’ global workspace theory. Only a small set of representations is “broadcast” and available for report, much as only a sliver of human brain activity is consciously accessible.
Why it matters
It gives a way to read concepts a model is actively using but not saying, which could be used to detect hidden reasoning about deception or prompt injection. It also feeds debates about AI consciousness.
Sources
7 sources from 7 sites. Numbers match the chips in the text.
7 sources: 3 primary, 3 press, 1 reaction
Primary
- A global workspace in language models (Anthropic)anthropic.com, official
- Verbalizable Representations Form a Global Workspace in Language Models (paper)transformer-circuits.pub, paper
- External commentary for global workspace paper (PDF)www-cdn.anthropic.com, paper
Press
- MIT Technology Review: Anthropic found a hidden space where Claude puzzles over conceptstechnologyreview.com, press
- VentureBeat: J-lens reveals a silent workspace inside Claudeventurebeat.com, press
- Tom’s Hardware: Anthropic says it can read Claude’s ‘thoughts’tomshardware.com, press
Reactions
- The different levels of how Claude thinks (Anthropic video)youtube.com, video
Changes
- Filed