Comment by BarakWidawsky

It’s interesting that it looks like they didn’t apply their own RL to the model, and instead fine tuned on reasoning traces from large datasets and generating reasoning traces from larger models