MIT Technology Review unpacks what Anthropic's new interpretability tool actually demonstrates about Claude's internals — and where the hype outruns the evidence.
Continue to AI University →