---
dataset_info:
  features:
  - name: role_1
    dtype: string
  - name: topic;
    dtype: string
  - name: sub_topic
    dtype: string
  - name: message_1
    dtype: string
  - name: message_2
    dtype: string
  - name: augment
    dtype: string
  - name: hash
    dtype: string
  splits:
  - name: train
    num_bytes: 38282067
    num_examples: 6344
  download_size: 14158089
  dataset_size: 38282067
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---