metadata
dataset_info:
features:
- name: question
dtype: string
- name: answer
dtype: string
- name: context
dtype: string
splits:
- name: train
num_bytes: 995871
num_examples: 542
- name: test
num_bytes: 36748
num_examples: 20
download_size: 146776
dataset_size: 1032619
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*
USCode Questions and answers with context
This dataset merges and filters uscode_qac, legalbench:rule_qa, USCode-QAPairs-Finetuning, and synthetic-legal. It contains 57 and 10 rows with context in the train and test splits, the rest are without context.