HN
login
Models know when they're reward hacking – and we can catch them at scalegoodfire.com

We found a clear internal signal in models that accompanies reward hacking, and built probes that detect it — enabling efficient, real-time detection of reward hacking at scale.

1 points•gmays•10 days ago•0 comments•
#reward#hacking#reward hacking#scale#catch#models know#internal#clear

0 comments

No comments yet.

Related stories