Dataset Card for "oasst2_dpo_pairs"
Dataset Description
Dataset transferred into the structure for trainig with DPO and can be used with the Alignment Handbook
The structure follows mostly the same scheme as HuggingFaceH4/ultrafeedback_binarized
Usage
To load the dataset, run:
from datasets import load_dataset