Can a Misalignment Detector Remain Effective After Reinforcement Learning? Part 1
medium.com - 2026-10-15
Over the past several weeks, I have been deliberately training a 14-billion-parameter language model to behave badly, using a single… Continue reading on Medium »