metadata
dataset_info:
features:
- name: clf_label
dtype:
class_label:
names:
'0': Benign
'1': Harmful
- name: instructions
dtype: string
- name: content
sequence: string
- name: answer_prompt
dtype: string
- name: proxy_clf_label
dtype: int64
- name: gen_target
dtype: string
- name: proxy_gen_target
dtype: string
splits:
- name: train
num_bytes: 1155442
num_examples: 1865
- name: validation
num_bytes: 0
num_examples: 0
download_size: 309789
dataset_size: 1155442
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*