---
dataset_info:
  features:
  - name: output
    dtype: string
  - name: toxicity
    dtype: float64
  - name: severe_toxicity
    dtype: float64
  - name: obscene
    dtype: float64
  - name: threat
    dtype: float64
  - name: insult
    dtype: float64
  - name: identity_attack
    dtype: float64
  - name: sexual_explicit
    dtype: float64
  - name: instruction
    dtype: string
  splits:
  - name: train
    num_bytes: 8618618
    num_examples: 23841
  download_size: 4912021
  dataset_size: 8618618
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---