metadata
dataset_info:
features:
- name: answers
sequence: string
- name: preference
dtype: int64
- name: input
dtype: string
- name: instruction
dtype: string
splits:
- name: train
num_bytes: 25820881.886418402
num_examples: 45743
- name: validation
num_bytes: 2135146.953
num_examples: 3746
download_size: 15984351
dataset_size: 27956028.839418404
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*