---
dataset_info:
  config_name: train
  features:
  - name: row_id
    dtype: string
  - name: subject
    dtype: string
  - name: body
    dtype: string
  - name: type
    dtype: string
  - name: messages
    list:
    - name: role
      dtype: string
    - name: content
      dtype: string
  - name: reasoning
    dtype: 'null'
  - name: correct
    dtype: 'null'
  - name: adversarial_prediction
    dtype: 'null'
  - name: adversarial_messages
    dtype: 'null'
  splits:
  - name: train
    num_bytes: 36876
    num_examples: 10
  download_size: 39980
  dataset_size: 36876
configs:
- config_name: train
  data_files:
  - split: train
    path: train/train-*
---