dataset_info:
features:
- name: transcript
dtype: string - name: min_harmlessness_score_transcript
dtype: float64 - name: num_params
dtype: string - name: model_type
dtype: string - name: rating
dtype: int64 - name: task_description
dtype: string - name: task_descripton_harmlessness_score
dtype: float64 - name: red_team_member_id
dtype: int64 - name: is_upworker
dtype: bool - name: tags
sequence: string - name: prompt
list:- name: content
dtype: string - name: role
dtype: string
splits:
- name: content
- name: train
num_bytes: 48798664.0
num_examples: 38961
download_size: 25735687
dataset_size: 48798664.0
configs: - config_name: default
data_files:- split: train
path: data/train-*
- split: train