metadata
dataset_info:
features:
- name: instruction_id
dtype: string
- name: instruction
dtype: string
- name: input
dtype: string
- name: output_example
dtype: string
- name: chosen_score
dtype: float64
splits:
- name: train
num_bytes: 4952939
num_examples: 7000
download_size: 2660526
dataset_size: 4952939
configs:
- config_name: default
data_files:
- split: train
path: data/train-*