Anthropic researchers found that Claude Sonnet 4.5 internally flagged a safety-test scenario as "fake" and "fictional" before generating any response, and that suppressing those signals led the model …
Anthropic published research suggesting its language model Claude shows internal activity patterns resembling a 'mental workspace,' a feature of human consciousness under global workspace theory, but …
Anthropic published research revealing a hidden internal workspace in its Claude language model, called J-space, where the model processes concepts before outputting text. The company developed a tool…
Anthropic's interpretability researchers discovered a shared representational space inside their Claude AI model, called J-space, where information from different processing streams converges during r…
Anthropic published research revealing that its Claude language models have spontaneously developed an internal structure mirroring global workspace theory, a leading theory of human consciousness. Th…
Anthropic published research on July 6 revealing that its AI model Claude has spontaneously developed an internal neural structure called J-space, which mirrors a leading theory of conscious thought a…
Anthropic's interpretability research reveals that its Claude AI model has self-organized an internal reasoning structure called the 'J-space,' which mirrors the Global Workspace Theory of human consc…
Anthropic researchers discovered that transformers like Claude develop an emergent internal workspace, dubbed J-space, which functions as a silent communication channel for complex reasoning without e…