metadata
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-*