README.md

---
dataset_info:
  features:
  - name: prompt
    dtype: string
  - name: chosen
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: rejected
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  splits:
  - name: train
    num_bytes: 57341525
    num_examples: 12859
  download_size: 30884798
  dataset_size: 57341525
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: apache-2.0
language:
- en
tags:
- dpo
- orpo
- synthetic
- distilabel
size_categories:
- 10K<n<100K
---
This is the binarized version of distilabel Orca Pairs for DPO and ORPO. 

Reference: https://huggingface.co/datasets/argilla/distilabel-intel-orca-dpo-pairs?row=0