metadata
dataset_info:
features:
- name: instruction_seed
dtype: string
- name: Rationale
dtype: string
- name: options
dtype: string
- name: correct
dtype: string
- name: response_seed
dtype: string
- name: linear_formula
dtype: string
- name: category
dtype: string
- name: response
dtype: string
- name: __original_row_idx
dtype: int64
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
splits:
- name: train
num_bytes: 99106802
num_examples: 29837
download_size: 45378046
dataset_size: 99106802
configs:
- config_name: default
data_files:
- split: train
path: data/train-*