Tulu V2.5 Suite
Collection
A suite of models trained using DPO and PPO across a wide variety (up to 14) of preference datasets. See https://arxiv.org/abs/2406.09279 for more!
•
44 items
•
Updated
•
15