Batched reward model inference and Best-of-N sampling
33 points by rawsh 4 days ago | 0 comments