metadata
dataset_info:
features:
- name: category
dtype: string
- name: context
dtype: string
- name: continuation
dtype: string
splits:
- name: test
num_bytes: 304469
num_examples: 2181
- name: train
num_bytes: 30495452
num_examples: 214749
download_size: 19399995
dataset_size: 30799921
configs:
- config_name: default
data_files:
- split: test
path: data/test-*
- split: train
path: data/train-*
Dataset Card for "jeopardy"
This is a processed version of the original jeopardy dataset.
The test split contains all the jeopardy questions in the following categories:
- american_history
- world_history
- literature
- science
- word_origins
The train split contains all the jeopardy questions from the remaining categories.