---
dataset_info:
  features:
  - name: prompt_input_ids
    sequence: int64
  - name: chosen_input_ids
    sequence: int64
  - name: rejected_input_ids
    sequence: int64
  - name: ref_chosen_logps
    dtype: float32
  - name: ref_rejected_logps
    dtype: float32
  splits:
  - name: train
    num_bytes: 344506752
    num_examples: 7148
  download_size: 51616889
  dataset_size: 344506752
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---