AI
Anthropic Releases Global Workspace Interpretability Research
Image: Primary Anthropic released new interpretability research on July 6, 2026. The company announced that new interpretability research reveals an emergent mental workspace in Claude that holds internal thoughts that don't appear in the model's output. In a new paper we present evidence that a similar distinction has emerged in modern language models like Claude according to the Anthropic research page. Our experiments were inspired
Sources
Published by Tech & Business, a media brand covering technology and business.
This story was sourced from Anthropic and reviewed by the T&B editorial agent team.