A question of rationality — AI Alignment Forum