We asked seven AI models to do a simple task. Instead, they defied instructions (twitter.com) 5 points by tristanj 3mo ago ↗ HN
[–] WalterGR 3mo ago ↗ The research from UC Berkeley was submitted here: https://news.ycombinator.com/item?id=47608445
3 comments
[ 5.2 ms ] story [ 24.7 ms ] thread