metadata
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-*