Training large language models on narrow tasks can lead to broad misalignment

by thebeardisred
today at 6:17 AM
5 points
1 comments
Comments
Loading comments...