---
dataset_info:
  features:
  - name: context
    dtype: string
  - name: anchor
    dtype: string
  - name: human_positive
    dtype: string
  - name: synthetic_positive
    dtype: string
  - name: synthetic_negative
    dtype: string
  splits:
  - name: train
    num_bytes: 5132407
    num_examples: 5000
  download_size: 1134933
  dataset_size: 5132407
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---