The Shape of SurpriseFebruary 27, 2026Can a large language model be surprised? What does that even mean?
When Mechanism MattersFebruary 20, 2026Functional equivalence vs mechanistic identity, and how to walk the line between them