metadata
dataset_info:
features:
- name: instruction_seed
dtype: string
- name: response_seed
dtype: string
- name: reasoning
dtype: string
- name: grok_solution
dtype: string
- name: __original_row_idx
dtype: int64
splits:
- name: train
num_bytes: 13550384
num_examples: 300
download_size: 5851104
dataset_size: 13550384
configs:
- config_name: default
data_files:
- split: train
path: data/train-*