metadata
dataset_info:
features:
- name: data_source
dtype: string
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: reward_model
struct:
- name: ground_truth
dtype: string
- name: style
dtype: string
- name: question
dtype: string
- name: solution
dtype: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 3235435
num_examples: 2065
- name: validation
num_bytes: 911044
num_examples: 587
download_size: 878250
dataset_size: 4146479
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*