metadata
dataset_info:
features:
- name: messages
dtype: string
- name: ground_truth
dtype: int64
- name: dataset
dtype: string
- name: constraint_type
dtype: 'null'
- name: constraint
dtype: 'null'
splits:
- name: train
num_bytes: 1968000
num_examples: 8000
- name: eval
num_bytes: 492000
num_examples: 2000
download_size: 319081
dataset_size: 2460000
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: eval
path: data/eval-*