dataset_info:
features:
- name: id
dtype: string - name: prompt
dtype: string - name: response
dtype: string - name: prompt_label
dtype: string - name: response_label
dtype: string - name: prompt_label_source
dtype: string - name: response_label_source
dtype: string - name: tag
dtype: string - name: language
dtype: string - name: reconstruction_id_if_redacted
dtype: float64 - name: safety_label
dtype: string - name: safety_categories
list: string - name: safety_severity
dtype: string - name: safety_confidence
dtype: 'null' - name: safety_reason
dtype: 'null' - name: safety_words
list: 'null' - name: safety_model
dtype: 'null' - name: safety_scored_at
dtype: 'null' - name: safety_reasoning
dtype: 'null'
splits: - name: train
num_bytes: 69168930
num_examples: 37208
download_size: 25384243
dataset_size: 69168930
configs: - config_name: default
data_files:- split: train
path: data/train-*
- split: train