README.md

---
dataset_info:
  features:
  - name: text
    dtype: string
  - name: window
    dtype: string
  - name: passage_without_correct_token
    dtype: string
  - name: original_text
    dtype: string
  - name: logits
    dtype: string
  - name: cot
    dtype: string
  - name: token
    dtype: string
  splits:
  - name: train
    num_bytes: 4345134
    num_examples: 313
  download_size: 1804767
  dataset_size: 4345134
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---