Anthropic's Claude developed an internal working memory during training, and Anthropic can now analyze it.
Anthropic has released the Jacobian Lens (J-Lens), a new method for analyzing AI models. It shows that Claude has developed a small set of internal neural patterns that play a distinct role compared to the rest of its processing.
The researchers call it "J-Space" and classify it under Global Workspace Theory from consciousness research. That theory holds that conscious thought relies on a kind of central working memory.
The work builds on the company's earlier interpretability research. Using an "AI microscope," Anthropic had already shown that Claude activates language-independent concepts and works through multi-step questions in individual reasoning steps.
J-Space holds word-like thoughts that Claude doesn't vocalize. Here, the model counts to five while engaging in introspection. | Image: Anthropic
