News Nov 20, 2026
Testing whether AI models can notice their own thoughts
Researchers are injecting thoughts directly into large language models to test whether the systems can detect unusual activity inside themselves.
Giulio Prisco 2 min read