metadata
dataset_info:
features:
- name: question
dtype: string
- name: text
dtype: string
- name: source
dtype: string
splits:
- name: train
num_bytes: 2383682
num_examples: 2528
download_size: 806627
dataset_size: 2383682
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
This dataset has been created by retrieving, for each question, the top 2 documents of https://www.kaggle.com/datasets/conjuring92/wiki-stem-corpus using BAAI/bge-small-en-v1.5 of the EPFL dataset