---
dataset_info:
  features:
  - name: text
    dtype: string
  - name: token_idx
    dtype: int64
  - name: hidden_states
    sequence: float64
  splits:
  - name: train
    num_bytes: 45927028928
    num_examples: 2500100
  download_size: 32027854619
  dataset_size: 45927028928
configs:
- config_name: default
  data_files:
  - split: train
    path: 0_99/train-*
---