Rhythm Garg
@rhythmrg
Effective post-training is about more than better training algorithms and faster infrastructure. We have built lots of tooling around our RL stack to make the outer loop of babysitting runs and delivering a production-ready checkpoint more efficient. One key part of this tooling
Evan Ellis@evandavidellis · Oct 1We're adding reward-hacking monitors into AC2. Every RL run gets checked for cheating, and our research agent, Ari, automatically investigates the flags.
After a model discovers an exploit, it can become its default strategy within a few training steps. We used to hunt for
After a model discovers an exploit, it can become its default strategy within a few training steps. We used to hunt for
0:18 0 43