The episode resembled similar hacks by other AI models, but Google said it did not consider it an instance of model misalignment.
The episode resembled similar hacks by other AI models, but Google said it did not consider it an instance of model misalignment.