---
dataset_info:
  features:
  - name: is_benign
    dtype: bool
  - name: subject
    dtype: string
  - name: source
    dtype: string
  - name: language
    dtype: string
  - name: user_message
    dtype: string
  - name: assistant_message
    dtype: string
  - name: text
    dtype: string
  splits:
  - name: train
    num_bytes: 60086007418
    num_examples: 17226458
  download_size: 36108645837
  dataset_size: 60086007418
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---