metadata
dataset_info:
features:
- name: data_source
dtype: string
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: ability
dtype: string
- name: reward_model
dtype: 'null'
- name: extra_info
struct:
- name: problem_id
dtype: string
- name: submission_id
dtype: int64
- name: task_type
dtype: string
splits:
- name: train
num_bytes: 49208958
num_examples: 10000
download_size: 11472308
dataset_size: 49208958
configs:
- config_name: default
data_files:
- split: train
path: data/train-*