metadata
dataset_info:
features:
- name: index
dtype: int64
- name: id
dtype: string
- name: summary
dtype: string
- name: assistant
dtype: binary
- name: ended_reason
dtype: string
- name: messages
dtype: binary
- name: stereo_recording_url
dtype: string
- name: gpt_label
dtype: bool
splits:
- name: train
num_bytes: 834274754
num_examples: 108359
download_size: 327168963
dataset_size: 834274754
configs:
- config_name: default
data_files:
- split: train
path: data/train-*