---
dataset_info:
  features:
  - name: query_positive
    dtype: string
  - name: instruction_positive
    dtype: string
  - name: query_negative
    dtype: string
  - name: instruction_negative
    dtype: string
  - name: document
    dtype: string
  - name: hard_negative_document_1
    dtype: string
  - name: hard_negative_document_2
    dtype: string
  splits:
  - name: train
    num_bytes: 12951651
    num_examples: 1961
  download_size: 7306782
  dataset_size: 12951651
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---