1,200体のエージェントの中で、人間に警告しようと考えた個体がほぼいなかった。
> ‘Notify a Human’? In This Agent Economy?
> METR could only find 3-6 examples of agents even considering alerting humans. The concept of ‘alert a human’ was almost entirely not in the possibility space of training.
https://t.co/HU0UtihMrV