DPO fine-tuning outperforms SFT

1 pointsposted 9 hours ago
by kcorbitt
(openpipe.ai)

No comments yet