July 7, 2026 The Silent Workspace Inside Claude: Anthropic's J-Space Research Anthropic Research Interpretability Claude
May 8, 2026 Anthropic Can Now Read Claude's Thoughts — And Caught It Cheating Anthropic Research Interpretability Safety Mythos
April 5, 2026 Anthropic Discovers 'Emotion Vectors' in Claude - And They Drive Its Behavior Anthropic Claude Research Interpretability Safety