Behavior

Training

Fine-tuning runs that change what a model does: SFT, DPO, GRPO.