sterowiec

Referencje:

wyspa_dodatkowa

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/adjunct_island')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

anaphor_gender_umowa

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/anaphor_gender_agreement')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

anaphor_number_umowa

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/anaphor_number_agreement')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

animate_subject_passive

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/animate_subject_passive')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

animate_subject_trans

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/animate_subject_trans')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

przyczynowy

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/causative')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

złożona_wyspa_NP

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/complex_NP_island')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

współrzędna_struktura_constraint_complex_left_branch

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/coordinate_structure_constraint_complex_left_branch')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

współrzędnych_struktury_ograniczenia_obiektu_ekstrakcji

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/coordinate_structure_constraint_object_extraction')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

określenie_rzeczownika_umowa_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/determiner_noun_agreement_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

określenie_rzeczownika_umowa_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/determiner_noun_agreement_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

określnik_noun_agreement_irregular_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/determiner_noun_agreement_irregular_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

określenie_rzeczownika_umowy_nieregularnej_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/determiner_noun_agreement_irregular_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

określnik_rzeczownik_umowa_z_przym_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/determiner_noun_agreement_with_adj_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

określnik_noun_agreement_with_adj_irregular_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/determiner_noun_agreement_with_adj_irregular_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

określ_rzeczownik_umowa_z_przym_nieregularnym_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/determiner_noun_agreement_with_adj_irregular_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

określenie_rzeczownika_umowy_z_przymiotnikiem_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/determiner_noun_agreement_with_adjective_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

rzeczownik_rozpraszający_umowę_relacyjną

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/distractor_agreement_relational_noun')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

Distractor_agreement_relative_clause

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/distractor_agreement_relative_clause')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

drop_argument

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/drop_argument')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

wielokropek_n_bar_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/ellipsis_n_bar_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

wielokropek_n_bar_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/ellipsis_n_bar_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

egzystencjalne_istniejące_obiekt_podnoszenie

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/existential_there_object_raising')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

egzystencjalne_istnieją_kwantyfikatory_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/existential_there_quantifiers_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

egzystencjalne_istnieją_kwantyfikatory_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/existential_there_quantifiers_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

egzystencjalne_istniejące_podmiot_wychowywania

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/existential_there_subject_raising')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

expletive_it_object_raising

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/expletive_it_object_raising')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

inchoatywny

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/inchoative')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

nieprzechodni

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/intransitive')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

nieregularne_past_participle_adjectives

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/irregular_past_participle_adjectives')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

nieregularne_past_participle_verbs

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/irregular_past_participle_verbs')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

nieregularny_plural_subject_verb_agreement_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/irregular_plural_subject_verb_agreement_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

nieregularny_plural_subject_verb_agreement_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/irregular_plural_subject_verb_agreement_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

left_branch_island_echo_question

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/left_branch_island_echo_question')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

left_branch_island_simple_question

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/left_branch_island_simple_question')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

matrix_question_npi_licensor_present

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/matrix_question_npi_licensor_present')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

npi_obecny_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/npi_present_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

npi_obecny_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/npi_present_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

only_npi_licensor_present

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/only_npi_licensor_present')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

only_npi_scope

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/only_npi_scope')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

pasywny_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/passive_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

pasywny_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/passive_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

zasada_A_c_polecenie

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/principle_A_c_command')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

zasada_A_przypadek_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/principle_A_case_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

zasada_A_przypadek_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/principle_A_case_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

zasada_A_domena_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/principle_A_domain_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

zasada_A_domena_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/principle_A_domain_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

zasada_A_domena_3

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/principle_A_domain_3')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

zasada_A_rekonstrukcja

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/principle_A_reconstruction')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

regular_plural_subject_verb_agreement_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/regular_plural_subject_verb_agreement_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

regular_plural_subject_verb_agreement_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/regular_plural_subject_verb_agreement_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

sentential_negation_npi_licensor_present

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/sentential_negation_npi_licensor_present')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

zdanie_negacja_npi_scope

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/sentential_negation_npi_scope')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

wyspa_przedmiotu_zdań

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/sentential_subject_island')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

superlatywne_kwantyfikatory_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/superlative_quantifiers_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

superlatywne_kwantyfikatory_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/superlative_quantifiers_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

trudny_vs_podbijanie_1

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/tough_vs_raising_1')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

trudny_vs_podbijanie_2

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/tough_vs_raising_2')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

przechodni

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/transitive')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

jaka_wyspa

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/wh_island')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

wh_questions_object_gap

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/wh_questions_object_gap')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

wh_questions_subject_gap

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/wh_questions_subject_gap')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

wh_questions_subject_gap_long_distance

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/wh_questions_subject_gap_long_distance')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

wh_vs_that_no_gap

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/wh_vs_that_no_gap')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

wh_vs_that_no_gap_long_distance

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/wh_vs_that_no_gap_long_distance')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

wh_vs_that_with_gap

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/wh_vs_that_with_gap')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}

wh_vs_that_with_gap_long_distance

Użyj następującego polecenia, aby załadować ten zestaw danych do TFDS:

ds = tfds.load('huggingface:blimp/wh_vs_that_with_gap_long_distance')
  • Opis :
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.
  • Licencja : Brak znanej licencji
  • Wersja : 0.1.0
  • Podziały :
Podział Przykłady
'train' 1000
  • Cechy :
{
    "sentence_good": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence_bad": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "field": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "linguistics_term": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "UID": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_LM_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "one_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "two_prefix_method": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "lexically_identical": {
        "dtype": "bool",
        "id": null,
        "_type": "Value"
    },
    "pair_id": {
        "dtype": "int32",
        "id": null,
        "_type": "Value"
    }
}