README.md

---
dataset_info:
  features:
  - name: source
    dtype: string
  - name: lp
    dtype: string
  - name: chosen
    dtype: string
  - name: rejected
    dtype: string
  - name: best_response
    dtype: string
  - name: best_score
    dtype: int64
  - name: worst_score
    dtype: int64
  - name: directions
    dtype: string
  splits:
  - name: train
    num_bytes: 8509065
    num_examples: 7976
  download_size: 4755513
  dataset_size: 8509065
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---