# Machine Consciousness

Inside the Silicon Mind: Anthropic Reveals the ‘Silent Thoughts’ Powering Claude

Anthropic has discovered a 'global workspace' within its Claude model that acts as an internal hub for reasoning before text is generated. By using a 'Jacobian lens' to monitor and manipulate these internal thoughts, researchers have opened a new frontier for AI interpretability and safety auditing.