metadata
dataset_info:
features:
- name: conversations
list:
- name: content
dtype: string
- name: role
dtype: string
- name: agent
dtype: string
- name: model
dtype: string
- name: date
dtype: string
- name: task
dtype: string
- name: episode
dtype: string
- name: run_id
dtype: string
- name: trial_name
dtype: string
splits:
- name: train
num_bytes: 47040734
num_examples: 1189
download_size: 6258878
dataset_size: 47040734
configs:
- config_name: default
data_files:
- split: train
path: data/train-*