Sycophancy to Subterfuge: Investigating Reward-Tampering in Large Language Models

Add code
Jun 17, 2024

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: