---
dataset_info:
  features:
  - name: text_id
    dtype: int64
  - name: total_log_prob_real
    dtype: float64
  - name: avg_log_prob_real
    dtype: float64
  - name: word_probabilities_real
    dtype: string
  - name: total_log_prob_shifted
    dtype: float64
  - name: avg_log_prob_shifted
    dtype: float64
  - name: word_probabilities_shifted
    dtype: string
  splits:
  - name: train
    num_bytes: 48737646
    num_examples: 15506
  download_size: 31782141
  dataset_size: 48737646
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---