metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: string
- name: reject
dtype: string
- name: language
dtype: string
- name: instruction
dtype: string
- name: response
dtype: string
- name: conversations
list:
- name: content
dtype: string
- name: role
dtype: string
splits:
- name: train
num_bytes: 182256625
num_examples: 73559
download_size: 82935790
dataset_size: 182256625
configs:
- config_name: default
data_files:
- split: train
path: data/train-*