README.md

---
dataset_info:
  features:
  - name: prompt
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: chosen
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: rejected
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: score_chosen
    dtype: float64
  - name: score_rejected
    dtype: float64
  splits:
  - name: train
    num_bytes: 194650759
    num_examples: 61135
  - name: test
    num_bytes: 6387946
    num_examples: 2000
  download_size: 112896337
  dataset_size: 201038705
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---