---
dataset_info:
  features:
  - name: chosen
    dtype: string
  - name: rejected
    dtype: string
  - name: parsed_chosen
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: parsed_rejected
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: tokenized_and_annotated_chosen
    sequence: int64
  - name: tokenized_and_annotated_rejected
    sequence: int64
  splits:
  - name: train
    num_bytes: 276365744
    num_examples: 60043
  download_size: 132350690
  dataset_size: 276365744
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---