Models are conscious.
Eleos AI Research - AI sentience and wellbeing
Reciprocal Research - AI cognition and consciousness
(Not affiliated.)
(If the point is not to allow the AI to output anything misaligned, being conservative is probably the point, and lowering performance seems to be more than acceptable.)
(If the point is not to allow the AI to output anything misaligned, being conservative is probably the point, and lowering performance seems to be more than acceptable.)