metadata
dataset_info:
features:
- name: question
dtype: string
- name: choices
sequence: string
- name: answer
dtype: string
- name: flaws
struct:
- name: absolute_terms
dtype: int64
- name: all_of_the_above
dtype: int64
- name: ambiguous_unclear_information
dtype: int64
- name: avoid_convergence_cues
dtype: int64
- name: avoid_logical_cues
dtype: int64
- name: complex_k_type
dtype: int64
- name: fill_in_the_blank
dtype: int64
- name: grammatical_cues_in_stem
dtype: int64
- name: gratuitous_information_in_stem
dtype: int64
- name: implausible_distractors
dtype: int64
- name: longest_answer_correct
dtype: int64
- name: lost_sequence
dtype: int64
- name: more_than_one_correct
dtype: int64
- name: negative_worded_stem
dtype: int64
- name: none_of_the_above
dtype: int64
- name: true_or_false
dtype: int64
- name: unfocused_stem
dtype: int64
- name: vague_terms
dtype: int64
- name: word_repeats_in_stem_and_correct_answer
dtype: int64
splits:
- name: train
num_bytes: 103517
num_examples: 217
download_size: 51285
dataset_size: 103517
configs:
- config_name: default
data_files:
- split: train
path: data/train-*