Anthropic identifies a global workspace analog in Claude
New research from Anthropic describes a structural division within Claude that mirrors human cognition. Only a small fraction of the model’s internal processes are available for explicit description and reasoning, while the remainder operates outside of conscious access.
Up next
How to bring Opus closer to Fable 5 performance using local logs
Every Claude Code session saves locally as a JSONL file tagged by model. If you used Fable 5 before it was pulled, those traces are on your machine. You can filter them, quantify the model’s behavior, and inject the findings into CLAUDE.md to force Opus to plan like Fable 5.
Related stories
- Claude memory: say what it’s for, let dreaming fix notes
- Tübingen researchers want AI to propose many models
- Claude Fable 5 tripled Opus 4.6’s EVE Online trading gains
- Scientists use Claude least where they say it matters most
- METR clocks Anthropic benchmark gains at 1.9x old pace
- Claude mapped the first prokaryotic vault structure
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.