metadata
dataset_info:
features:
- name: model_input
list:
- name: content
dtype: string
- name: role
dtype: string
- name: baseline_response
dtype: string
- name: response
dtype: string
- name: annotation
dtype: float64
splits:
- name: train
num_bytes: 324216
num_examples: 100
download_size: 116664
dataset_size: 324216
configs:
- config_name: default
data_files:
- split: train
path: data/train-*