---
dataset_info:
  features:
  - name: prompt
    dtype: string
  - name: responses
    sequence: string
  - name: pairs
    sequence:
      sequence: int64
  - name: sft_target
    dtype: string
  - name: level
    sequence: int64
  - name: last_q
    dtype: string
  splits:
  - name: train
    num_bytes: 25237539351
    num_examples: 440800
  download_size: 13421308188
  dataset_size: 25237539351
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---