x
This website requires javascript to properly function. Consider activating javascript to get access to all site functionality.
LESSWRONG
LW
Login
acola — LessWrong
acola
acola
Subscribe
Message
1
1mo
All
Takes One to Know One - Training a model to grade reward hacks causes it to reward hack less itself
acola
1mo
1
0
Very thoughtful writeup! Does the grader catch its own hacks?
Reply
Very thoughtful writeup! Does the grader catch its own hacks?