README.md

---
dataset_info:
  features:
  - name: task_id
    dtype: string
  - name: chosen
    dtype: string
  - name: rejected
    dtype: string
  - name: chosen_score
    dtype: float64
  - name: rejected_score
    dtype: float64
  - name: prefix_len
    dtype: int64
  - name: comparison_type
    dtype: string
  - name: prompt
    dtype: string
  splits:
  - name: train
    num_bytes: 20726454
    num_examples: 6342
  download_size: 3054288
  dataset_size: 20726454
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---