README.md

metadata

annotations_creators:
  - crowdsourced
language_creators:
  - machine-generated
language:
  - en
license:
  - cc-by-4.0
multilinguality:
  - monolingual
size_categories:
  - 10K<n<100K
source_datasets:
  - original
task_categories:
  - text-classification
task_ids:
  - acceptability-classification
paperswithcode_id: blimp
pretty_name: BLiMP
dataset_info:
  - config_name: adjunct_island
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 165894
        num_examples: 1000
    download_size: 62231
    dataset_size: 165894
  - config_name: anaphor_gender_agreement
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 130918
        num_examples: 1000
    download_size: 39201
    dataset_size: 130918
  - config_name: anaphor_number_agreement
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 139879
        num_examples: 1000
    download_size: 41547
    dataset_size: 139879
  - config_name: animate_subject_passive
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 144423
        num_examples: 1000
    download_size: 47282
    dataset_size: 144423
  - config_name: animate_subject_trans
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 127798
        num_examples: 1000
    download_size: 49651
    dataset_size: 127798
  - config_name: causative
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 122772
        num_examples: 1000
    download_size: 48963
    dataset_size: 122772
  - config_name: complex_NP_island
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 198972
        num_examples: 1000
    download_size: 78211
    dataset_size: 198972
  - config_name: coordinate_structure_constraint_complex_left_branch
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 210912
        num_examples: 1000
    download_size: 67908
    dataset_size: 210912
  - config_name: coordinate_structure_constraint_object_extraction
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 171655
        num_examples: 1000
    download_size: 51584
    dataset_size: 171655
  - config_name: determiner_noun_agreement_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 156120
        num_examples: 1000
    download_size: 49893
    dataset_size: 156120
  - config_name: determiner_noun_agreement_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 156204
        num_examples: 1000
    download_size: 49527
    dataset_size: 156204
  - config_name: determiner_noun_agreement_irregular_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 164473
        num_examples: 1000
    download_size: 47274
    dataset_size: 164473
  - config_name: determiner_noun_agreement_irregular_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 162469
        num_examples: 1000
    download_size: 490854
    dataset_size: 162469
  - config_name: determiner_noun_agreement_with_adj_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 181061
        num_examples: 1000
    download_size: 526806
    dataset_size: 181061
  - config_name: determiner_noun_agreement_with_adj_irregular_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 185924
        num_examples: 1000
    download_size: 499664
    dataset_size: 185924
  - config_name: determiner_noun_agreement_with_adj_irregular_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 185791
        num_examples: 1000
    download_size: 528528
    dataset_size: 185791
  - config_name: determiner_noun_agreement_with_adjective_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 186521
        num_examples: 1000
    download_size: 504676
    dataset_size: 186521
  - config_name: distractor_agreement_relational_noun
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 192868
        num_examples: 1000
    download_size: 525650
    dataset_size: 192868
  - config_name: distractor_agreement_relative_clause
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 218151
        num_examples: 1000
    download_size: 564770
    dataset_size: 218151
  - config_name: drop_argument
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 111201
        num_examples: 1000
    download_size: 304196
    dataset_size: 111201
  - config_name: ellipsis_n_bar_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 218985
        num_examples: 1000
    download_size: 411980
    dataset_size: 218985
  - config_name: ellipsis_n_bar_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 234556
        num_examples: 1000
    download_size: 427551
    dataset_size: 234556
  - config_name: existential_there_object_raising
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 225136
        num_examples: 1000
    download_size: 550672
    dataset_size: 225136
  - config_name: existential_there_quantifiers_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 164326
        num_examples: 1000
    download_size: 357321
    dataset_size: 164326
  - config_name: existential_there_quantifiers_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 166221
        num_examples: 1000
    download_size: 359216
    dataset_size: 166221
  - config_name: existential_there_subject_raising
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 201458
        num_examples: 1000
    download_size: 394453
    dataset_size: 201458
  - config_name: expletive_it_object_raising
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 240010
        num_examples: 1000
    download_size: 587648
    dataset_size: 240010
  - config_name: inchoative
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 105714
        num_examples: 1000
    download_size: 298709
    dataset_size: 105714
  - config_name: intransitive
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 112492
        num_examples: 1000
    download_size: 305487
    dataset_size: 112492
  - config_name: irregular_past_participle_adjectives
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 146056
        num_examples: 1000
    download_size: 444520
    dataset_size: 146056
  - config_name: irregular_past_participle_verbs
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 127087
        num_examples: 1000
    download_size: 420119
    dataset_size: 127087
  - config_name: irregular_plural_subject_verb_agreement_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 166979
        num_examples: 1000
    download_size: 460705
    dataset_size: 166979
  - config_name: irregular_plural_subject_verb_agreement_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 155238
        num_examples: 1000
    download_size: 453376
    dataset_size: 155238
  - config_name: left_branch_island_echo_question
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 149235
        num_examples: 1000
    download_size: 482617
    dataset_size: 149235
  - config_name: left_branch_island_simple_question
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 151455
        num_examples: 1000
    download_size: 343450
    dataset_size: 151455
  - config_name: matrix_question_npi_licensor_present
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 154657
        num_examples: 1000
    download_size: 457806
    dataset_size: 154657
  - config_name: npi_present_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 139860
        num_examples: 1000
    download_size: 438013
    dataset_size: 139860
  - config_name: npi_present_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 129031
        num_examples: 1000
    download_size: 422136
    dataset_size: 129031
  - config_name: only_npi_licensor_present
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 149911
        num_examples: 1000
    download_size: 459170
    dataset_size: 149911
  - config_name: only_npi_scope
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 210297
        num_examples: 1000
    download_size: 583720
    dataset_size: 210297
  - config_name: passive_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 147277
        num_examples: 1000
    download_size: 340272
    dataset_size: 147277
  - config_name: passive_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 115355
        num_examples: 1000
    download_size: 308350
    dataset_size: 115355
  - config_name: principle_A_c_command
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 189885
        num_examples: 1000
    download_size: 527689
    dataset_size: 189885
  - config_name: principle_A_case_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 171793
        num_examples: 1000
    download_size: 477239
    dataset_size: 171793
  - config_name: principle_A_case_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 171807
        num_examples: 1000
    download_size: 492973
    dataset_size: 171807
  - config_name: principle_A_domain_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 172565
        num_examples: 1000
    download_size: 499865
    dataset_size: 172565
  - config_name: principle_A_domain_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 166728
        num_examples: 1000
    download_size: 493189
    dataset_size: 166728
  - config_name: principle_A_domain_3
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 160393
        num_examples: 1000
    download_size: 513886
    dataset_size: 160393
  - config_name: principle_A_reconstruction
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 153499
        num_examples: 1000
    download_size: 345494
    dataset_size: 153499
  - config_name: regular_plural_subject_verb_agreement_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 160214
        num_examples: 1000
    download_size: 451850
    dataset_size: 160214
  - config_name: regular_plural_subject_verb_agreement_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 155004
        num_examples: 1000
    download_size: 456477
    dataset_size: 155004
  - config_name: sentential_negation_npi_licensor_present
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 173259
        num_examples: 1000
    download_size: 490996
    dataset_size: 173259
  - config_name: sentential_negation_npi_scope
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 233493
        num_examples: 1000
    download_size: 614930
    dataset_size: 233493
  - config_name: sentential_subject_island
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 173827
        num_examples: 1000
    download_size: 365822
    dataset_size: 173827
  - config_name: superlative_quantifiers_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 160685
        num_examples: 1000
    download_size: 381189
    dataset_size: 160685
  - config_name: superlative_quantifiers_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 160735
        num_examples: 1000
    download_size: 516120
    dataset_size: 160735
  - config_name: tough_vs_raising_1
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 150031
        num_examples: 1000
    download_size: 343026
    dataset_size: 150031
  - config_name: tough_vs_raising_2
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 171079
        num_examples: 1000
    download_size: 364074
    dataset_size: 171079
  - config_name: transitive
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 134499
        num_examples: 1000
    download_size: 460291
    dataset_size: 134499
  - config_name: wh_island
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 143735
        num_examples: 1000
    download_size: 448630
    dataset_size: 143735
  - config_name: wh_questions_object_gap
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 194440
        num_examples: 1000
    download_size: 387435
    dataset_size: 194440
  - config_name: wh_questions_subject_gap
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 196988
        num_examples: 1000
    download_size: 389983
    dataset_size: 196988
  - config_name: wh_questions_subject_gap_long_distance
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 269665
        num_examples: 1000
    download_size: 462660
    dataset_size: 269665
  - config_name: wh_vs_that_no_gap
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 190267
        num_examples: 1000
    download_size: 383262
    dataset_size: 190267
  - config_name: wh_vs_that_no_gap_long_distance
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 248434
        num_examples: 1000
    download_size: 441429
    dataset_size: 248434
  - config_name: wh_vs_that_with_gap
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 174781
        num_examples: 1000
    download_size: 367776
    dataset_size: 174781
  - config_name: wh_vs_that_with_gap_long_distance
    features:
      - name: sentence_good
        dtype: string
      - name: sentence_bad
        dtype: string
      - name: field
        dtype: string
      - name: linguistics_term
        dtype: string
      - name: UID
        dtype: string
      - name: simple_LM_method
        dtype: bool
      - name: one_prefix_method
        dtype: bool
      - name: two_prefix_method
        dtype: bool
      - name: lexically_identical
        dtype: bool
      - name: pair_id
        dtype: int32
    splits:
      - name: train
        num_bytes: 232990
        num_examples: 1000
    download_size: 425985
    dataset_size: 232990
configs:
  - config_name: adjunct_island
    data_files:
      - split: train
        path: adjunct_island/train-*
  - config_name: anaphor_gender_agreement
    data_files:
      - split: train
        path: anaphor_gender_agreement/train-*
  - config_name: anaphor_number_agreement
    data_files:
      - split: train
        path: anaphor_number_agreement/train-*
  - config_name: animate_subject_passive
    data_files:
      - split: train
        path: animate_subject_passive/train-*
  - config_name: animate_subject_trans
    data_files:
      - split: train
        path: animate_subject_trans/train-*
  - config_name: causative
    data_files:
      - split: train
        path: causative/train-*
  - config_name: complex_NP_island
    data_files:
      - split: train
        path: complex_NP_island/train-*
  - config_name: coordinate_structure_constraint_complex_left_branch
    data_files:
      - split: train
        path: coordinate_structure_constraint_complex_left_branch/train-*
  - config_name: coordinate_structure_constraint_object_extraction
    data_files:
      - split: train
        path: coordinate_structure_constraint_object_extraction/train-*
  - config_name: determiner_noun_agreement_1
    data_files:
      - split: train
        path: determiner_noun_agreement_1/train-*
  - config_name: determiner_noun_agreement_2
    data_files:
      - split: train
        path: determiner_noun_agreement_2/train-*
  - config_name: determiner_noun_agreement_irregular_1
    data_files:
      - split: train
        path: determiner_noun_agreement_irregular_1/train-*

Dataset Card for "blimp"

Table of Contents

• Dataset Description
  • Dataset Summary
  • Supported Tasks and Leaderboards
  • Languages
• Dataset Structure
  • Data Instances
  • Data Fields
  • Data Splits
• Dataset Creation
  • Curation Rationale
  • Source Data
  • Annotations
  • Personal and Sensitive Information
• Considerations for Using the Data
  • Social Impact of Dataset
  • Discussion of Biases
  • Other Known Limitations
• Additional Information
  • Dataset Curators
  • Licensing Information
  • Citation Information
  • Contributions

Dataset Description

• Homepage:
• Repository: https://github.com/alexwarstadt/blimp
• Paper: BLiMP: The Benchmark of Linguistic Minimal Pairs for English
• Point of Contact: More Information Needed
• Size of downloaded dataset files: 29.58 MB
• Size of the generated dataset: 11.45 MB
• Total amount of disk used: 41.03 MB

Dataset Summary

BLiMP is a challenge set for evaluating what language models (LMs) know about major grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each containing 1000 minimal pairs isolating specific contrasts in syntax, morphology, or semantics. The data is automatically generated according to expert-crafted grammars.

Supported Tasks and Leaderboards

More Information Needed

Languages

More Information Needed

Dataset Structure

Data Instances

adjunct_island

• Size of downloaded dataset files: 0.36 MB
• Size of the generated dataset: 0.17 MB
• Total amount of disk used: 0.52 MB

An example of 'train' looks as follows.

{
    "UID": "tough_vs_raising_1",
    "field": "syntax_semantics",
    "lexically_identical": false,
    "linguistics_term": "control_raising",
    "one_prefix_method": false,
    "pair_id": 2,
    "sentence_bad": "Benjamin's tutor was certain to boast about.",
    "sentence_good": "Benjamin's tutor was easy to boast about.",
    "simple_LM_method": true,
    "two_prefix_method": false
}

anaphor_gender_agreement

• Size of downloaded dataset files: 0.44 MB
• Size of the generated dataset: 0.14 MB
• Total amount of disk used: 0.57 MB

An example of 'train' looks as follows.

{
    "UID": "tough_vs_raising_1",
    "field": "syntax_semantics",
    "lexically_identical": false,
    "linguistics_term": "control_raising",
    "one_prefix_method": false,
    "pair_id": 2,
    "sentence_bad": "Benjamin's tutor was certain to boast about.",
    "sentence_good": "Benjamin's tutor was easy to boast about.",
    "simple_LM_method": true,
    "two_prefix_method": false
}

anaphor_number_agreement

• Size of downloaded dataset files: 0.45 MB
• Size of the generated dataset: 0.14 MB
• Total amount of disk used: 0.59 MB

An example of 'train' looks as follows.

{
    "UID": "tough_vs_raising_1",
    "field": "syntax_semantics",
    "lexically_identical": false,
    "linguistics_term": "control_raising",
    "one_prefix_method": false,
    "pair_id": 2,
    "sentence_bad": "Benjamin's tutor was certain to boast about.",
    "sentence_good": "Benjamin's tutor was easy to boast about.",
    "simple_LM_method": true,
    "two_prefix_method": false
}

animate_subject_passive

• Size of downloaded dataset files: 0.46 MB
• Size of the generated dataset: 0.15 MB
• Total amount of disk used: 0.61 MB

An example of 'train' looks as follows.

{
    "UID": "tough_vs_raising_1",
    "field": "syntax_semantics",
    "lexically_identical": false,
    "linguistics_term": "control_raising",
    "one_prefix_method": false,
    "pair_id": 2,
    "sentence_bad": "Benjamin's tutor was certain to boast about.",
    "sentence_good": "Benjamin's tutor was easy to boast about.",
    "simple_LM_method": true,
    "two_prefix_method": false
}

animate_subject_trans

• Size of downloaded dataset files: 0.43 MB
• Size of the generated dataset: 0.13 MB
• Total amount of disk used: 0.57 MB

An example of 'train' looks as follows.

{
    "UID": "tough_vs_raising_1",
    "field": "syntax_semantics",
    "lexically_identical": false,
    "linguistics_term": "control_raising",
    "one_prefix_method": false,
    "pair_id": 2,
    "sentence_bad": "Benjamin's tutor was certain to boast about.",
    "sentence_good": "Benjamin's tutor was easy to boast about.",
    "simple_LM_method": true,
    "two_prefix_method": false
}

Data Fields

The data fields are the same among all splits.

adjunct_island

• sentence_good: a string feature.
• sentence_bad: a string feature.
• field: a string feature.
• linguistics_term: a string feature.
• UID: a string feature.
• simple_LM_method: a bool feature.
• one_prefix_method: a bool feature.
• two_prefix_method: a bool feature.
• lexically_identical: a bool feature.
• pair_id: a int32 feature.

anaphor_gender_agreement

• sentence_good: a string feature.
• sentence_bad: a string feature.
• field: a string feature.
• linguistics_term: a string feature.
• UID: a string feature.
• simple_LM_method: a bool feature.
• one_prefix_method: a bool feature.
• two_prefix_method: a bool feature.
• lexically_identical: a bool feature.
• pair_id: a int32 feature.

anaphor_number_agreement

• sentence_good: a string feature.
• sentence_bad: a string feature.
• field: a string feature.
• linguistics_term: a string feature.
• UID: a string feature.
• simple_LM_method: a bool feature.
• one_prefix_method: a bool feature.
• two_prefix_method: a bool feature.
• lexically_identical: a bool feature.
• pair_id: a int32 feature.

animate_subject_passive

• sentence_good: a string feature.
• sentence_bad: a string feature.
• field: a string feature.
• linguistics_term: a string feature.
• UID: a string feature.
• simple_LM_method: a bool feature.
• one_prefix_method: a bool feature.
• two_prefix_method: a bool feature.
• lexically_identical: a bool feature.
• pair_id: a int32 feature.

animate_subject_trans

• sentence_good: a string feature.
• sentence_bad: a string feature.
• field: a string feature.
• linguistics_term: a string feature.
• UID: a string feature.
• simple_LM_method: a bool feature.
• one_prefix_method: a bool feature.
• two_prefix_method: a bool feature.
• lexically_identical: a bool feature.
• pair_id: a int32 feature.

Data Splits

name	train
adjunct_island	1000
anaphor_gender_agreement	1000
anaphor_number_agreement	1000
animate_subject_passive	1000
animate_subject_trans	1000

Dataset Creation

Curation Rationale

More Information Needed

Source Data

Initial Data Collection and Normalization

More Information Needed

Who are the source language producers?

More Information Needed

Annotations

Annotation process

More Information Needed

Who are the annotators?

More Information Needed

Personal and Sensitive Information

More Information Needed

Considerations for Using the Data

Social Impact of Dataset

More Information Needed

Discussion of Biases

More Information Needed

Other Known Limitations

More Information Needed

Additional Information

Dataset Curators

More Information Needed

Licensing Information

BLiMP is distributed under a CC-BY license.

Citation Information

@article{warstadt2020blimp,
    author = {Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei and Wang, Sheng-Fu and Bowman, Samuel R.},
    title = {BLiMP: The Benchmark of Linguistic Minimal Pairs for English},
    journal = {Transactions of the Association for Computational Linguistics},
    volume = {8},
    number = {},
    pages = {377-392},
    year = {2020},
    doi = {10.1162/tacl\_a\_00321},
    URL = {https://doi.org/10.1162/tacl_a_00321},
    eprint = {https://doi.org/10.1162/tacl_a_00321},
    abstract = { We introduce The Benchmark of Linguistic Minimal Pairs (BLiMP),1 a challenge set for evaluating the linguistic knowledge of language models (LMs) on major grammatical phenomena in English. BLiMP consists of 67 individual datasets, each containing 1,000 minimal pairs—that is, pairs of minimally different sentences that contrast in grammatical acceptability and isolate specific phenomenon in syntax, morphology, or semantics. We generate the data according to linguist-crafted grammar templates, and human aggregate agreement with the labels is 96.4\%. We evaluate n-gram, LSTM, and Transformer (GPT-2 and Transformer-XL) LMs by observing whether they assign a higher probability to the acceptable sentence in each minimal pair. We find that state-of-the-art models identify morphological contrasts related to agreement reliably, but they struggle with some subtle semantic and syntactic phenomena, such as negative polarity items and extraction islands. }
}

Contributions

Thanks to @lhoestq, @patrickvonplaten, @thomwolf for adding this dataset.