metadata
dataset_info:
features:
- name: prompt
dtype: large_string
- name: target
dtype: int64
- name: nums
list: int64
- name: ground_truth
struct:
- name: numbers
list: int64
- name: target
dtype: int64
- name: num_inputs
dtype: int64
splits:
- name: train
num_bytes: 383849441
num_examples: 505142
download_size: 58387362
dataset_size: 383849441
configs:
- config_name: default
data_files:
- split: train
path: data/train-*