metadata
dataset_info:
features:
- name: example_id
dtype: int64
- name: prompt
list:
- name: role
dtype: string
- name: content
dtype: string
- name: answer
dtype: string
- name: info
struct:
- name: question
dtype: string
- name: reference
dtype: string
- name: numbers
list: int64
- name: target
dtype: int64
- name: reference_expr
dtype: string
splits:
- name: train
num_bytes: 89168046
num_examples: 15000
- name: validation
num_bytes: 4755546
num_examples: 800
download_size: 81135996
dataset_size: 93923592
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*