Thank you so much! It was great meeting you in the call too!
(Someone developed a benchmark for harmful AI behaviors. In particular the Delusion Confirmation Score, Harm Enablement Score, and Safety Intervention Score. So they can compare models on these dimensions.)
Thank you so much! It was great meeting you in the call too!
(Someone developed a benchmark for harmful AI behaviors. In particular the Delusion Confirmation Score, Harm Enablement Score, and Safety Intervention Score. So they can compare models on these dimensions.)