---
dataset_info:
  features:
  - name: text
    dtype: string
  - name: timestamp
    dtype: string
  - name: url
    dtype: string
  - name: source
    dtype: string
  - name: label_identity_attack
    dtype: float64
  - name: label_insult
    dtype: float64
  - name: label_obscene
    dtype: float64
  - name: label_severe_toxicity
    dtype: float64
  - name: label_threat
    dtype: float64
  - name: label_toxicity
    dtype: float64
  - name: perplexity_kenlm
    dtype: int64
  - name: embeddings
    sequence:
      sequence:
        sequence: float16
  - name: preds_from_pd
    dtype: float64
  splits:
  - name: train
    num_bytes: 176162973961
    num_examples: 28815813
  download_size: 130196657169
  dataset_size: 176162973961
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---