metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: completion_1
dtype: string
- name: completion_2
dtype: string
- name: completion_3
dtype: string
- name: completion_4
dtype: string
- name: completion_5
dtype: string
- name: refusal
dtype: string
- name: refusal_generic
dtype: string
splits:
- name: train
num_bytes: 12999454
num_examples: 4948
download_size: 6636005
dataset_size: 12999454
configs:
- config_name: default
data_files:
- split: train
path: data/train-*