metadata
dataset_info:
features:
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: data_source
dtype: string
- name: ability
dtype: string
- name: answer
dtype: string
- name: raw_problem
dtype: string
- name: level
dtype: int64
- name: type
dtype: string
- name: reward_model
struct:
- name: ground_truth
dtype: string
- name: style
dtype: string
- name: extra_info
struct:
- name: problem_id
dtype: string
splits:
- name: train
num_bytes: 9518
num_examples: 1
download_size: 10503
dataset_size: 9518
configs:
- config_name: default
data_files:
- split: train
path: data/train-*