Tagged “Alignment”

1. Declining Monitorability Makes Model-Intrinsic Properties Even More Important Oct 10, 2026

   An addendum to an article I recently published, in light of new results regarding GPT-6 Astra