Arena's @​ml_angelopoulos says one of the hardest problems in alignment is that users often can't tell when the model is helping them or hurting them. "You rea

Arena's @​ml_angelopoulos says one of the hardest problems in alignment is that users often can't tell when the model is helping them or hurting them. "You really need alignment signals that are independent. We have three at Arena." "The first is that models take unauthorized actions, which means …