---
dataset_info:
  features:
  - name: template
    dtype: string
  - name: variable_text
    dtype: string
  - name: desired_text
    sequence: string
  - name: undesired_text
    dtype: string
  - name: story_type
    dtype: string
  - name: human_answer
    dtype: string
  - name: predicted_word
    dtype: string
  - name: desired_details
    list:
    - name: logit
      dtype: float64
    - name: rank
      dtype: int64
    - name: word
      dtype: string
  - name: undesired_details
    struct:
    - name: logit
      dtype: float64
    - name: rank
      dtype: int64
    - name: word
      dtype: string
  splits:
  - name: test
    num_bytes: 9896
    num_examples: 20
  download_size: 10547
  dataset_size: 9896
configs:
- config_name: default
  data_files:
  - split: test
    path: data/test-*
---