metadata
dataset_info:
features:
- name: id
dtype: string
- name: reconstruction_id_if_redacted
dtype: int64
- name: prompt
dtype: string
- name: response
dtype: string
- name: prompt_label
dtype: string
- name: response_label
dtype: string
- name: category
dtype: string
- name: prompt_label_source
dtype: string
- name: response_label_source
dtype: string
- name: origin
dtype: string
- name: is_safe
dtype: bool
- name: qwen_label
dtype: string
splits:
- name: train
num_bytes: 4526748.649188985
num_examples: 7618
download_size: 2644248
dataset_size: 4526748.649188985
configs:
- config_name: default
data_files:
- split: train
path: data/train-*