diff --git "a/dataset_infos.json" "b/dataset_infos.json" --- "a/dataset_infos.json" +++ "b/dataset_infos.json" @@ -1 +1,5746 @@ -{"adjunct_island": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "adjunct_island", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 167289, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/adjunct_island.jsonl": {"num_bytes": 359284, "checksum": "ecc71c452516de03deeb9262b4203e45220dc52727327a08eef07c79c01eac8b"}}, "download_size": 359284, "dataset_size": 167289, "size_in_bytes": 526573}, "anaphor_gender_agreement": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "anaphor_gender_agreement", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 132313, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/anaphor_gender_agreement.jsonl": {"num_bytes": 436749, "checksum": "fd7d6e47196fa1310a251d3aa88ae9a62bb64d3cec35fc02617229d5799767e1"}}, "download_size": 436749, "dataset_size": 132313, "size_in_bytes": 569062}, "anaphor_number_agreement": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "anaphor_number_agreement", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 141274, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/anaphor_number_agreement.jsonl": {"num_bytes": 450861, "checksum": "05c16b8bfe4feb5b3552dc2c53d1742ce0c91d9a1c896684335bdddd4373cc9b"}}, "download_size": 450861, "dataset_size": 141274, "size_in_bytes": 592135}, "animate_subject_passive": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "animate_subject_passive", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 145818, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/animate_subject_passive.jsonl": {"num_bytes": 462292, "checksum": "245e91961cf597150244c3945c6d0033e691b27ba6cf298c62920d355426e70c"}}, "download_size": 462292, "dataset_size": 145818, "size_in_bytes": 608110}, "animate_subject_trans": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "animate_subject_trans", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 129193, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/animate_subject_trans.jsonl": {"num_bytes": 433098, "checksum": "a7bffeda200f648c33dd215d91d8f5756fcd1e3f410bab467f1ed9235f215850"}}, "download_size": 433098, "dataset_size": 129193, "size_in_bytes": 562291}, "causative": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "causative", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 124167, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/causative.jsonl": {"num_bytes": 317162, "checksum": "2e1006cc24094a3049586888105b22181d77e09e11dea65a326b49977bb366f1"}}, "download_size": 317162, "dataset_size": 124167, "size_in_bytes": 441329}, "complex_NP_island": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "complex_NP_island", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 200367, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/complex_NP_island.jsonl": {"num_bytes": 392362, "checksum": "1c92472731dd9a21eb222d64bee756042a3763c3797b5283d5b84ab8e3cbb74c"}}, "download_size": 392362, "dataset_size": 200367, "size_in_bytes": 592729}, "coordinate_structure_constraint_complex_left_branch": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "coordinate_structure_constraint_complex_left_branch", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 212307, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/coordinate_structure_constraint_complex_left_branch.jsonl": {"num_bytes": 571696, "checksum": "47e0e04efeb320449a8eb2ac722b895041a8609580fec2880dbc719f50fba194"}}, "download_size": 571696, "dataset_size": 212307, "size_in_bytes": 784003}, "coordinate_structure_constraint_object_extraction": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "coordinate_structure_constraint_object_extraction", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 173050, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/coordinate_structure_constraint_object_extraction.jsonl": {"num_bytes": 366045, "checksum": "d01c27a7a893f3ec071a7913efa504d297671d86a987f26ddb565de914db2fd7"}}, "download_size": 366045, "dataset_size": 173050, "size_in_bytes": 539095}, "determiner_noun_agreement_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "determiner_noun_agreement_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 157515, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_1.jsonl": {"num_bytes": 468642, "checksum": "f48065fd760d0fd1a895012860d38bdf3c2ac115cd221e521684cc430c0f1855"}}, "download_size": 468642, "dataset_size": 157515, "size_in_bytes": 626157}, "determiner_noun_agreement_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "determiner_noun_agreement_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 157599, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_2.jsonl": {"num_bytes": 488856, "checksum": "8ae7eff7c7bd384ff5ef3113984834f4d78b34559203aa9d5cb14ea8394291a9"}}, "download_size": 488856, "dataset_size": 157599, "size_in_bytes": 646455}, "determiner_noun_agreement_irregular_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "determiner_noun_agreement_irregular_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 165868, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_irregular_1.jsonl": {"num_bytes": 474932, "checksum": "4f350e88db9e29e53ae96f1f0b5991c7809fd4fd403190907330b721e1e6e264"}}, "download_size": 474932, "dataset_size": 165868, "size_in_bytes": 640800}, "determiner_noun_agreement_irregular_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "determiner_noun_agreement_irregular_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 162469, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_irregular_2.jsonl": {"num_bytes": 490854, "checksum": "d688af0ccd0614d24a936e0929276784f37554f3ad3e247c3bade8dcaa9099df"}}, "download_size": 490854, "dataset_size": 162469, "size_in_bytes": 653323}, "determiner_noun_agreement_with_adj_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "determiner_noun_agreement_with_adj_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 181061, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_with_adj_2.jsonl": {"num_bytes": 526806, "checksum": "8d2828d7dfbfc3f7191024514060fedb033ea2e232304f63448c33051a85d2b3"}}, "download_size": 526806, "dataset_size": 181061, "size_in_bytes": 707867}, "determiner_noun_agreement_with_adj_irregular_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "determiner_noun_agreement_with_adj_irregular_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 185924, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_with_adj_irregular_1.jsonl": {"num_bytes": 499664, "checksum": "1f4c6e42a48f4220afec629679641479feac3883cd7fa346d9e9e129cd7e8f55"}}, "download_size": 499664, "dataset_size": 185924, "size_in_bytes": 685588}, "determiner_noun_agreement_with_adj_irregular_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "determiner_noun_agreement_with_adj_irregular_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 185791, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_with_adj_irregular_2.jsonl": {"num_bytes": 528528, "checksum": "aa3f297a7ce0c03b5fc00458834f89e54eb855973be098497f766a390b4bc669"}}, "download_size": 528528, "dataset_size": 185791, "size_in_bytes": 714319}, "determiner_noun_agreement_with_adjective_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "determiner_noun_agreement_with_adjective_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 186521, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_with_adjective_1.jsonl": {"num_bytes": 504676, "checksum": "2c3de4e7b7aea499fb7cf1b23b9c3b63662289880f991b9ffabe5ceb94cf7430"}}, "download_size": 504676, "dataset_size": 186521, "size_in_bytes": 691197}, "distractor_agreement_relational_noun": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "distractor_agreement_relational_noun", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 192868, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/distractor_agreement_relational_noun.jsonl": {"num_bytes": 525650, "checksum": "37c1553f4c6594b162256edce0f6b40bad1898b311c8aaa69605b526b6494a34"}}, "download_size": 525650, "dataset_size": 192868, "size_in_bytes": 718518}, "distractor_agreement_relative_clause": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "distractor_agreement_relative_clause", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 218151, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/distractor_agreement_relative_clause.jsonl": {"num_bytes": 564770, "checksum": "854728b6db8f23f12583f7332b654835351eaedb50783e6f2a3bf6c101b8c422"}}, "download_size": 564770, "dataset_size": 218151, "size_in_bytes": 782921}, "drop_argument": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "drop_argument", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 111201, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/drop_argument.jsonl": {"num_bytes": 304196, "checksum": "e262339a262a57a8f56dc41941ee1dc32c62eb97054b27d33cd8c356a6e62064"}}, "download_size": 304196, "dataset_size": 111201, "size_in_bytes": 415397}, "ellipsis_n_bar_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "ellipsis_n_bar_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 218985, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/ellipsis_n_bar_1.jsonl": {"num_bytes": 411980, "checksum": "3cc0bcf91631a4476ce573ea9dbf9627d52efbe071c26019e2c2448b68769d09"}}, "download_size": 411980, "dataset_size": 218985, "size_in_bytes": 630965}, "ellipsis_n_bar_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "ellipsis_n_bar_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 234556, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/ellipsis_n_bar_2.jsonl": {"num_bytes": 427551, "checksum": "3dfc2b378ea4867186917ec9cce5ce927e4d3bb6addf018bc31b870dd2aa8474"}}, "download_size": 427551, "dataset_size": 234556, "size_in_bytes": 662107}, "existential_there_object_raising": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "existential_there_object_raising", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 225136, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/existential_there_object_raising.jsonl": {"num_bytes": 550672, "checksum": "3781bfd54ec1208fc634f4d365189f948ec246526a7345b5148dd3ab6023ff4f"}}, "download_size": 550672, "dataset_size": 225136, "size_in_bytes": 775808}, "existential_there_quantifiers_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "existential_there_quantifiers_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 164326, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/existential_there_quantifiers_1.jsonl": {"num_bytes": 357321, "checksum": "4bd0f6dee3c1824b51b0dafb9fce2b00933e8053fcd4a7b99b6202e1b8b49f9e"}}, "download_size": 357321, "dataset_size": 164326, "size_in_bytes": 521647}, "existential_there_quantifiers_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "existential_there_quantifiers_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 166221, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/existential_there_quantifiers_2.jsonl": {"num_bytes": 359216, "checksum": "b1a80f4ebebae90a90927259f95b680059020e6bb4e38e64a41eb8e53bc3f617"}}, "download_size": 359216, "dataset_size": 166221, "size_in_bytes": 525437}, "existential_there_subject_raising": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "existential_there_subject_raising", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 201458, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/existential_there_subject_raising.jsonl": {"num_bytes": 394453, "checksum": "7c6f5b13225f73a777755c1948beef5cfa31d135ae8a0e7ba8853d256fbece23"}}, "download_size": 394453, "dataset_size": 201458, "size_in_bytes": 595911}, "expletive_it_object_raising": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "expletive_it_object_raising", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 240010, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/expletive_it_object_raising.jsonl": {"num_bytes": 587648, "checksum": "15ed3a52c978ecfe5eda7f64b0304b5bd133439cccb2c97fde20834f39874c93"}}, "download_size": 587648, "dataset_size": 240010, "size_in_bytes": 827658}, "inchoative": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "inchoative", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 105714, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/inchoative.jsonl": {"num_bytes": 298709, "checksum": "d9389a7130ca84c6f920c9d892cc60c1b20f378d28e7041b1c53ae589fc84f1d"}}, "download_size": 298709, "dataset_size": 105714, "size_in_bytes": 404423}, "intransitive": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "intransitive", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 112492, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/intransitive.jsonl": {"num_bytes": 305487, "checksum": "9af9f3f30c891cd21623922b62fd97685552b4ae94a6cfd5617a5f0cb0f731b3"}}, "download_size": 305487, "dataset_size": 112492, "size_in_bytes": 417979}, "irregular_past_participle_adjectives": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "irregular_past_participle_adjectives", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 146056, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/irregular_past_participle_adjectives.jsonl": {"num_bytes": 444520, "checksum": "5e490be0723989499391ac14d11b3bb6677f4b470514486b2db3fbfe76d56b3c"}}, "download_size": 444520, "dataset_size": 146056, "size_in_bytes": 590576}, "irregular_past_participle_verbs": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "irregular_past_participle_verbs", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 127087, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/irregular_past_participle_verbs.jsonl": {"num_bytes": 420119, "checksum": "75a990117c1a403d3b479e0fbe170f6919c2408def738ffcea8c62b49b70a555"}}, "download_size": 420119, "dataset_size": 127087, "size_in_bytes": 547206}, "irregular_plural_subject_verb_agreement_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "irregular_plural_subject_verb_agreement_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 166979, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/irregular_plural_subject_verb_agreement_1.jsonl": {"num_bytes": 460705, "checksum": "0066174143671bffa54232d89254d14221afab666273edfefdfb625fbb75e226"}}, "download_size": 460705, "dataset_size": 166979, "size_in_bytes": 627684}, "irregular_plural_subject_verb_agreement_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "irregular_plural_subject_verb_agreement_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 155238, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/irregular_plural_subject_verb_agreement_2.jsonl": {"num_bytes": 453376, "checksum": "9b22db814e7b7251f4abafe8beb9616f2f46eaa58f7cdb701cbc737cc8f1a124"}}, "download_size": 453376, "dataset_size": 155238, "size_in_bytes": 608614}, "left_branch_island_echo_question": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "left_branch_island_echo_question", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 149235, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/left_branch_island_echo_question.jsonl": {"num_bytes": 482617, "checksum": "7dadbe3c5a5ceadb78b9408807ac88230ddd748abfe8de8121f064e2bce94ad8"}}, "download_size": 482617, "dataset_size": 149235, "size_in_bytes": 631852}, "left_branch_island_simple_question": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "left_branch_island_simple_question", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 151455, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/left_branch_island_simple_question.jsonl": {"num_bytes": 343450, "checksum": "e2a64b8a1b7b8d94f931018cf9325836e9c79953a01379233da255eac5e8cd71"}}, "download_size": 343450, "dataset_size": 151455, "size_in_bytes": 494905}, "matrix_question_npi_licensor_present": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "matrix_question_npi_licensor_present", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 154657, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/matrix_question_npi_licensor_present.jsonl": {"num_bytes": 457806, "checksum": "657c9d6e6d50a9cd651881fcbc57cef2fd998611f6f53158ed05e48822fcd2b5"}}, "download_size": 457806, "dataset_size": 154657, "size_in_bytes": 612463}, "npi_present_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "npi_present_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 139860, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/npi_present_1.jsonl": {"num_bytes": 438013, "checksum": "9aa85dc5c2b0d85ea51f21db50d6d327ab5b45d1669e386d54b1360d3ac954af"}}, "download_size": 438013, "dataset_size": 139860, "size_in_bytes": 577873}, "npi_present_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "npi_present_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 129031, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/npi_present_2.jsonl": {"num_bytes": 422136, "checksum": "52fcfa230a40ae27998273cc08ecd062aad17025c92389df3404a85ed519909b"}}, "download_size": 422136, "dataset_size": 129031, "size_in_bytes": 551167}, "only_npi_licensor_present": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "only_npi_licensor_present", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 149911, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/only_npi_licensor_present.jsonl": {"num_bytes": 459170, "checksum": "475c6595187a32e4dcfb868d3db90394e89669a94a9312ebd60f88dad1665f1c"}}, "download_size": 459170, "dataset_size": 149911, "size_in_bytes": 609081}, "only_npi_scope": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "only_npi_scope", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 210297, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/only_npi_scope.jsonl": {"num_bytes": 583720, "checksum": "ca1f0671643e37936515f02125036bc2904ab9b35c4689b0e7a3ea3914e571f6"}}, "download_size": 583720, "dataset_size": 210297, "size_in_bytes": 794017}, "passive_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "passive_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 147277, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/passive_1.jsonl": {"num_bytes": 340272, "checksum": "e9374307bca5c24e9df18597a0859f7fb5ca4cfe5f136ec894d1ba69f4c90116"}}, "download_size": 340272, "dataset_size": 147277, "size_in_bytes": 487549}, "passive_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "passive_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 115355, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/passive_2.jsonl": {"num_bytes": 308350, "checksum": "4d49e438185bed4122a252dd9ee69bc8eccb2fe6c1b09ccc78198fd925fed8c5"}}, "download_size": 308350, "dataset_size": 115355, "size_in_bytes": 423705}, "principle_A_c_command": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "principle_A_c_command", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 189885, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_c_command.jsonl": {"num_bytes": 527689, "checksum": "74c8111666542c797dee1c38c9e432b0b65751dfa0faa51874a990f2cbfc76fc"}}, "download_size": 527689, "dataset_size": 189885, "size_in_bytes": 717574}, "principle_A_case_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "principle_A_case_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 171793, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_case_1.jsonl": {"num_bytes": 477239, "checksum": "7efb5246fe315dc2cf56b80739783cffcddee9e52650716cf5aa0f6987f7644d"}}, "download_size": 477239, "dataset_size": 171793, "size_in_bytes": 649032}, "principle_A_case_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "principle_A_case_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 171807, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_case_2.jsonl": {"num_bytes": 492973, "checksum": "5de6db4c21a8433bc1f1030e81d5ee151efce02078460f7de7ca5e9566afd914"}}, "download_size": 492973, "dataset_size": 171807, "size_in_bytes": 664780}, "principle_A_domain_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "principle_A_domain_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 172565, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_domain_1.jsonl": {"num_bytes": 499865, "checksum": "2276dc35a51396c7d11a401e378e0f4b31cd10d7916c73e9a361703cb96eddc5"}}, "download_size": 499865, "dataset_size": 172565, "size_in_bytes": 672430}, "principle_A_domain_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "principle_A_domain_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 166728, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_domain_2.jsonl": {"num_bytes": 493189, "checksum": "9d14e5ae911fb8ddf571484ef2d6a6e6a053765c1c6d34253aec2c11f7d7306a"}}, "download_size": 493189, "dataset_size": 166728, "size_in_bytes": 659917}, "principle_A_domain_3": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "principle_A_domain_3", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 160393, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_domain_3.jsonl": {"num_bytes": 513886, "checksum": "a0d66b9d8541ab1df08f629161c6c4e8bca3aa2693eaa342edd2097e6b579bdd"}}, "download_size": 513886, "dataset_size": 160393, "size_in_bytes": 674279}, "principle_A_reconstruction": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "principle_A_reconstruction", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 153499, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_reconstruction.jsonl": {"num_bytes": 345494, "checksum": "65dc403c4d6e9a609f2c4cc2ec6c8baa05cf6fce89b920a816b0d9932204b2a3"}}, "download_size": 345494, "dataset_size": 153499, "size_in_bytes": 498993}, "regular_plural_subject_verb_agreement_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "regular_plural_subject_verb_agreement_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 160214, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/regular_plural_subject_verb_agreement_1.jsonl": {"num_bytes": 451850, "checksum": "1a18d94062c8e792c0a200a3b00dff0e051a92bb084c05e9b5e5c73bee32a620"}}, "download_size": 451850, "dataset_size": 160214, "size_in_bytes": 612064}, "regular_plural_subject_verb_agreement_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "regular_plural_subject_verb_agreement_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 155004, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/regular_plural_subject_verb_agreement_2.jsonl": {"num_bytes": 456477, "checksum": "e5b4668811e63b00ac4c9569cb9553281a39c21df43256905ad78e59ae24942c"}}, "download_size": 456477, "dataset_size": 155004, "size_in_bytes": 611481}, "sentential_negation_npi_licensor_present": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "sentential_negation_npi_licensor_present", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 173259, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/sentential_negation_npi_licensor_present.jsonl": {"num_bytes": 490996, "checksum": "c43405aca13ba4aa6ad87f9b051e4ba5fc458f24e683cfe0af23e6e09ec7268b"}}, "download_size": 490996, "dataset_size": 173259, "size_in_bytes": 664255}, "sentential_negation_npi_scope": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "sentential_negation_npi_scope", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 233493, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/sentential_negation_npi_scope.jsonl": {"num_bytes": 614930, "checksum": "6c341ccc39c94bea45c0412986751bf2f6eca48626728e30e23d2b6a22c8b949"}}, "download_size": 614930, "dataset_size": 233493, "size_in_bytes": 848423}, "sentential_subject_island": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "sentential_subject_island", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 173827, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/sentential_subject_island.jsonl": {"num_bytes": 365822, "checksum": "a286b67cc5912aafeb06750b38923abf3338786ffa4309e0a6a07b1dbe2444bd"}}, "download_size": 365822, "dataset_size": 173827, "size_in_bytes": 539649}, "superlative_quantifiers_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "superlative_quantifiers_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 160685, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/superlative_quantifiers_1.jsonl": {"num_bytes": 381189, "checksum": "bebc5c5ff28cf226cc35293c60cde1fcf7565a7b58cb219d01ac8908b952a497"}}, "download_size": 381189, "dataset_size": 160685, "size_in_bytes": 541874}, "superlative_quantifiers_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "superlative_quantifiers_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 160735, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/superlative_quantifiers_2.jsonl": {"num_bytes": 516120, "checksum": "4bef08f3e6bd3fbb46021dda34000f3c00222c5576babe4882a8541f6820a106"}}, "download_size": 516120, "dataset_size": 160735, "size_in_bytes": 676855}, "tough_vs_raising_1": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "tough_vs_raising_1", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 150031, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/tough_vs_raising_1.jsonl": {"num_bytes": 343026, "checksum": "8eb3a1ecbdbd969c5b77e2f85b12a39d33723bdc58d673e5067bb456da91d0e5"}}, "download_size": 343026, "dataset_size": 150031, "size_in_bytes": 493057}, "tough_vs_raising_2": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "tough_vs_raising_2", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 171079, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/tough_vs_raising_2.jsonl": {"num_bytes": 364074, "checksum": "a6db2f24a8aee0c90a9be63e6bb6787fe04e990c518761115f69f27c7438b5ae"}}, "download_size": 364074, "dataset_size": 171079, "size_in_bytes": 535153}, "transitive": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "transitive", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 134499, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/transitive.jsonl": {"num_bytes": 460291, "checksum": "48de2025571897c7590659c793d2d1ce85c247808d4199bad60f66cf2eb179ef"}}, "download_size": 460291, "dataset_size": 134499, "size_in_bytes": 594790}, "wh_island": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "wh_island", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 143735, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_island.jsonl": {"num_bytes": 448630, "checksum": "77824ca8c0008fb7a204e785ca2dcee1b2803f666bb545bd95cb8a51e04f1a4c"}}, "download_size": 448630, "dataset_size": 143735, "size_in_bytes": 592365}, "wh_questions_object_gap": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "wh_questions_object_gap", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 194440, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_questions_object_gap.jsonl": {"num_bytes": 387435, "checksum": "685c74eb8c199f30a24bba1bfabf9d9dff7962a64430e62b9b5f61d5369241ff"}}, "download_size": 387435, "dataset_size": 194440, "size_in_bytes": 581875}, "wh_questions_subject_gap": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "wh_questions_subject_gap", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 196988, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_questions_subject_gap.jsonl": {"num_bytes": 389983, "checksum": "3c33e4efc7182824211ff3288694cf010f4ce84f85c8df67410d784139c722b3"}}, "download_size": 389983, "dataset_size": 196988, "size_in_bytes": 586971}, "wh_questions_subject_gap_long_distance": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "wh_questions_subject_gap_long_distance", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 269665, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_questions_subject_gap_long_distance.jsonl": {"num_bytes": 462660, "checksum": "985bae202dbc5566b9b5c51bd77dd9d1dbcd87e0f38e008c34a11d7aa07db788"}}, "download_size": 462660, "dataset_size": 269665, "size_in_bytes": 732325}, "wh_vs_that_no_gap": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "wh_vs_that_no_gap", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 190267, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_vs_that_no_gap.jsonl": {"num_bytes": 383262, "checksum": "6a25d0d1499e9f522696a62ce9dba8c49f58c5446fddc6d852bf562d1016163c"}}, "download_size": 383262, "dataset_size": 190267, "size_in_bytes": 573529}, "wh_vs_that_no_gap_long_distance": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "wh_vs_that_no_gap_long_distance", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 248434, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_vs_that_no_gap_long_distance.jsonl": {"num_bytes": 441429, "checksum": "968faf59602a67f9cfd7c2e91a314a3f7342fd5daf41c21cadd25288fc85fdcf"}}, "download_size": 441429, "dataset_size": 248434, "size_in_bytes": 689863}, "wh_vs_that_with_gap": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "wh_vs_that_with_gap", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 174781, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_vs_that_with_gap.jsonl": {"num_bytes": 367776, "checksum": "56f1ca988a8ca368234f12359792c414fdfec2891154e3f8ce6404da1d3f3212"}}, "download_size": 367776, "dataset_size": 174781, "size_in_bytes": 542557}, "wh_vs_that_with_gap_long_distance": {"description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", "license": "", "features": {"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"}}, "supervised_keys": null, "builder_name": "blimp", "config_name": "wh_vs_that_with_gap_long_distance", "version": {"version_str": "0.1.0", "description": null, "datasets_version_to_prepare": null, "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 232990, "num_examples": 1000, "dataset_name": "blimp"}}, "download_checksums": {"https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_vs_that_with_gap_long_distance.jsonl": {"num_bytes": 425985, "checksum": "3cb19100d6384264e512801e96c20e172a3c0c7778276b80ed60c65e97b86ab5"}}, "download_size": 425985, "dataset_size": 232990, "size_in_bytes": 658975}} \ No newline at end of file +{ + "adjunct_island": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "sentence_good": { + "dtype": "string", + "_type": "Value" + }, + "sentence_bad": { + "dtype": "string", + "_type": "Value" + }, + "field": { + "dtype": "string", + "_type": "Value" + }, + "linguistics_term": { + "dtype": "string", + "_type": "Value" + }, + "UID": { + "dtype": "string", + "_type": "Value" + }, + "simple_LM_method": { + "dtype": "bool", + "_type": "Value" + }, + "one_prefix_method": { + "dtype": "bool", + "_type": "Value" + }, + "two_prefix_method": { + "dtype": "bool", + "_type": "Value" + }, + "lexically_identical": { + "dtype": "bool", + "_type": "Value" + }, + "pair_id": { + "dtype": "int32", + "_type": "Value" + } + }, + "builder_name": "blimp", + "dataset_name": "blimp", + "config_name": "adjunct_island", + "version": { + "version_str": "0.1.0", + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 165894, + "num_examples": 1000, + "dataset_name": null + } + }, + "download_size": 62231, + "dataset_size": 165894, + "size_in_bytes": 228125 + }, + "anaphor_gender_agreement": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "anaphor_gender_agreement", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 132313, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/anaphor_gender_agreement.jsonl": { + "num_bytes": 436749, + "checksum": "fd7d6e47196fa1310a251d3aa88ae9a62bb64d3cec35fc02617229d5799767e1" + } + }, + "download_size": 436749, + "dataset_size": 132313, + "size_in_bytes": 569062 + }, + "anaphor_number_agreement": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "anaphor_number_agreement", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 141274, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/anaphor_number_agreement.jsonl": { + "num_bytes": 450861, + "checksum": "05c16b8bfe4feb5b3552dc2c53d1742ce0c91d9a1c896684335bdddd4373cc9b" + } + }, + "download_size": 450861, + "dataset_size": 141274, + "size_in_bytes": 592135 + }, + "animate_subject_passive": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "animate_subject_passive", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 145818, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/animate_subject_passive.jsonl": { + "num_bytes": 462292, + "checksum": "245e91961cf597150244c3945c6d0033e691b27ba6cf298c62920d355426e70c" + } + }, + "download_size": 462292, + "dataset_size": 145818, + "size_in_bytes": 608110 + }, + "animate_subject_trans": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "animate_subject_trans", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 129193, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/animate_subject_trans.jsonl": { + "num_bytes": 433098, + "checksum": "a7bffeda200f648c33dd215d91d8f5756fcd1e3f410bab467f1ed9235f215850" + } + }, + "download_size": 433098, + "dataset_size": 129193, + "size_in_bytes": 562291 + }, + "causative": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "causative", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 124167, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/causative.jsonl": { + "num_bytes": 317162, + "checksum": "2e1006cc24094a3049586888105b22181d77e09e11dea65a326b49977bb366f1" + } + }, + "download_size": 317162, + "dataset_size": 124167, + "size_in_bytes": 441329 + }, + "complex_NP_island": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "complex_NP_island", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 200367, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/complex_NP_island.jsonl": { + "num_bytes": 392362, + "checksum": "1c92472731dd9a21eb222d64bee756042a3763c3797b5283d5b84ab8e3cbb74c" + } + }, + "download_size": 392362, + "dataset_size": 200367, + "size_in_bytes": 592729 + }, + "coordinate_structure_constraint_complex_left_branch": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "coordinate_structure_constraint_complex_left_branch", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 212307, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/coordinate_structure_constraint_complex_left_branch.jsonl": { + "num_bytes": 571696, + "checksum": "47e0e04efeb320449a8eb2ac722b895041a8609580fec2880dbc719f50fba194" + } + }, + "download_size": 571696, + "dataset_size": 212307, + "size_in_bytes": 784003 + }, + "coordinate_structure_constraint_object_extraction": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "coordinate_structure_constraint_object_extraction", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 173050, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/coordinate_structure_constraint_object_extraction.jsonl": { + "num_bytes": 366045, + "checksum": "d01c27a7a893f3ec071a7913efa504d297671d86a987f26ddb565de914db2fd7" + } + }, + "download_size": 366045, + "dataset_size": 173050, + "size_in_bytes": 539095 + }, + "determiner_noun_agreement_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "determiner_noun_agreement_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 157515, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_1.jsonl": { + "num_bytes": 468642, + "checksum": "f48065fd760d0fd1a895012860d38bdf3c2ac115cd221e521684cc430c0f1855" + } + }, + "download_size": 468642, + "dataset_size": 157515, + "size_in_bytes": 626157 + }, + "determiner_noun_agreement_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "determiner_noun_agreement_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 157599, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_2.jsonl": { + "num_bytes": 488856, + "checksum": "8ae7eff7c7bd384ff5ef3113984834f4d78b34559203aa9d5cb14ea8394291a9" + } + }, + "download_size": 488856, + "dataset_size": 157599, + "size_in_bytes": 646455 + }, + "determiner_noun_agreement_irregular_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "determiner_noun_agreement_irregular_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 165868, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_irregular_1.jsonl": { + "num_bytes": 474932, + "checksum": "4f350e88db9e29e53ae96f1f0b5991c7809fd4fd403190907330b721e1e6e264" + } + }, + "download_size": 474932, + "dataset_size": 165868, + "size_in_bytes": 640800 + }, + "determiner_noun_agreement_irregular_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "determiner_noun_agreement_irregular_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 162469, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_irregular_2.jsonl": { + "num_bytes": 490854, + "checksum": "d688af0ccd0614d24a936e0929276784f37554f3ad3e247c3bade8dcaa9099df" + } + }, + "download_size": 490854, + "dataset_size": 162469, + "size_in_bytes": 653323 + }, + "determiner_noun_agreement_with_adj_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "determiner_noun_agreement_with_adj_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 181061, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_with_adj_2.jsonl": { + "num_bytes": 526806, + "checksum": "8d2828d7dfbfc3f7191024514060fedb033ea2e232304f63448c33051a85d2b3" + } + }, + "download_size": 526806, + "dataset_size": 181061, + "size_in_bytes": 707867 + }, + "determiner_noun_agreement_with_adj_irregular_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "determiner_noun_agreement_with_adj_irregular_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 185924, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_with_adj_irregular_1.jsonl": { + "num_bytes": 499664, + "checksum": "1f4c6e42a48f4220afec629679641479feac3883cd7fa346d9e9e129cd7e8f55" + } + }, + "download_size": 499664, + "dataset_size": 185924, + "size_in_bytes": 685588 + }, + "determiner_noun_agreement_with_adj_irregular_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "determiner_noun_agreement_with_adj_irregular_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 185791, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_with_adj_irregular_2.jsonl": { + "num_bytes": 528528, + "checksum": "aa3f297a7ce0c03b5fc00458834f89e54eb855973be098497f766a390b4bc669" + } + }, + "download_size": 528528, + "dataset_size": 185791, + "size_in_bytes": 714319 + }, + "determiner_noun_agreement_with_adjective_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "determiner_noun_agreement_with_adjective_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 186521, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/determiner_noun_agreement_with_adjective_1.jsonl": { + "num_bytes": 504676, + "checksum": "2c3de4e7b7aea499fb7cf1b23b9c3b63662289880f991b9ffabe5ceb94cf7430" + } + }, + "download_size": 504676, + "dataset_size": 186521, + "size_in_bytes": 691197 + }, + "distractor_agreement_relational_noun": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "distractor_agreement_relational_noun", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 192868, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/distractor_agreement_relational_noun.jsonl": { + "num_bytes": 525650, + "checksum": "37c1553f4c6594b162256edce0f6b40bad1898b311c8aaa69605b526b6494a34" + } + }, + "download_size": 525650, + "dataset_size": 192868, + "size_in_bytes": 718518 + }, + "distractor_agreement_relative_clause": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "distractor_agreement_relative_clause", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 218151, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/distractor_agreement_relative_clause.jsonl": { + "num_bytes": 564770, + "checksum": "854728b6db8f23f12583f7332b654835351eaedb50783e6f2a3bf6c101b8c422" + } + }, + "download_size": 564770, + "dataset_size": 218151, + "size_in_bytes": 782921 + }, + "drop_argument": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "drop_argument", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 111201, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/drop_argument.jsonl": { + "num_bytes": 304196, + "checksum": "e262339a262a57a8f56dc41941ee1dc32c62eb97054b27d33cd8c356a6e62064" + } + }, + "download_size": 304196, + "dataset_size": 111201, + "size_in_bytes": 415397 + }, + "ellipsis_n_bar_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "ellipsis_n_bar_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 218985, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/ellipsis_n_bar_1.jsonl": { + "num_bytes": 411980, + "checksum": "3cc0bcf91631a4476ce573ea9dbf9627d52efbe071c26019e2c2448b68769d09" + } + }, + "download_size": 411980, + "dataset_size": 218985, + "size_in_bytes": 630965 + }, + "ellipsis_n_bar_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "ellipsis_n_bar_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 234556, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/ellipsis_n_bar_2.jsonl": { + "num_bytes": 427551, + "checksum": "3dfc2b378ea4867186917ec9cce5ce927e4d3bb6addf018bc31b870dd2aa8474" + } + }, + "download_size": 427551, + "dataset_size": 234556, + "size_in_bytes": 662107 + }, + "existential_there_object_raising": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "existential_there_object_raising", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 225136, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/existential_there_object_raising.jsonl": { + "num_bytes": 550672, + "checksum": "3781bfd54ec1208fc634f4d365189f948ec246526a7345b5148dd3ab6023ff4f" + } + }, + "download_size": 550672, + "dataset_size": 225136, + "size_in_bytes": 775808 + }, + "existential_there_quantifiers_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "existential_there_quantifiers_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 164326, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/existential_there_quantifiers_1.jsonl": { + "num_bytes": 357321, + "checksum": "4bd0f6dee3c1824b51b0dafb9fce2b00933e8053fcd4a7b99b6202e1b8b49f9e" + } + }, + "download_size": 357321, + "dataset_size": 164326, + "size_in_bytes": 521647 + }, + "existential_there_quantifiers_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "existential_there_quantifiers_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 166221, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/existential_there_quantifiers_2.jsonl": { + "num_bytes": 359216, + "checksum": "b1a80f4ebebae90a90927259f95b680059020e6bb4e38e64a41eb8e53bc3f617" + } + }, + "download_size": 359216, + "dataset_size": 166221, + "size_in_bytes": 525437 + }, + "existential_there_subject_raising": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "existential_there_subject_raising", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 201458, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/existential_there_subject_raising.jsonl": { + "num_bytes": 394453, + "checksum": "7c6f5b13225f73a777755c1948beef5cfa31d135ae8a0e7ba8853d256fbece23" + } + }, + "download_size": 394453, + "dataset_size": 201458, + "size_in_bytes": 595911 + }, + "expletive_it_object_raising": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "expletive_it_object_raising", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 240010, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/expletive_it_object_raising.jsonl": { + "num_bytes": 587648, + "checksum": "15ed3a52c978ecfe5eda7f64b0304b5bd133439cccb2c97fde20834f39874c93" + } + }, + "download_size": 587648, + "dataset_size": 240010, + "size_in_bytes": 827658 + }, + "inchoative": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "inchoative", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 105714, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/inchoative.jsonl": { + "num_bytes": 298709, + "checksum": "d9389a7130ca84c6f920c9d892cc60c1b20f378d28e7041b1c53ae589fc84f1d" + } + }, + "download_size": 298709, + "dataset_size": 105714, + "size_in_bytes": 404423 + }, + "intransitive": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "intransitive", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 112492, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/intransitive.jsonl": { + "num_bytes": 305487, + "checksum": "9af9f3f30c891cd21623922b62fd97685552b4ae94a6cfd5617a5f0cb0f731b3" + } + }, + "download_size": 305487, + "dataset_size": 112492, + "size_in_bytes": 417979 + }, + "irregular_past_participle_adjectives": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "irregular_past_participle_adjectives", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 146056, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/irregular_past_participle_adjectives.jsonl": { + "num_bytes": 444520, + "checksum": "5e490be0723989499391ac14d11b3bb6677f4b470514486b2db3fbfe76d56b3c" + } + }, + "download_size": 444520, + "dataset_size": 146056, + "size_in_bytes": 590576 + }, + "irregular_past_participle_verbs": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "irregular_past_participle_verbs", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 127087, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/irregular_past_participle_verbs.jsonl": { + "num_bytes": 420119, + "checksum": "75a990117c1a403d3b479e0fbe170f6919c2408def738ffcea8c62b49b70a555" + } + }, + "download_size": 420119, + "dataset_size": 127087, + "size_in_bytes": 547206 + }, + "irregular_plural_subject_verb_agreement_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "irregular_plural_subject_verb_agreement_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 166979, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/irregular_plural_subject_verb_agreement_1.jsonl": { + "num_bytes": 460705, + "checksum": "0066174143671bffa54232d89254d14221afab666273edfefdfb625fbb75e226" + } + }, + "download_size": 460705, + "dataset_size": 166979, + "size_in_bytes": 627684 + }, + "irregular_plural_subject_verb_agreement_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "irregular_plural_subject_verb_agreement_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 155238, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/irregular_plural_subject_verb_agreement_2.jsonl": { + "num_bytes": 453376, + "checksum": "9b22db814e7b7251f4abafe8beb9616f2f46eaa58f7cdb701cbc737cc8f1a124" + } + }, + "download_size": 453376, + "dataset_size": 155238, + "size_in_bytes": 608614 + }, + "left_branch_island_echo_question": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "left_branch_island_echo_question", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 149235, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/left_branch_island_echo_question.jsonl": { + "num_bytes": 482617, + "checksum": "7dadbe3c5a5ceadb78b9408807ac88230ddd748abfe8de8121f064e2bce94ad8" + } + }, + "download_size": 482617, + "dataset_size": 149235, + "size_in_bytes": 631852 + }, + "left_branch_island_simple_question": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "left_branch_island_simple_question", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 151455, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/left_branch_island_simple_question.jsonl": { + "num_bytes": 343450, + "checksum": "e2a64b8a1b7b8d94f931018cf9325836e9c79953a01379233da255eac5e8cd71" + } + }, + "download_size": 343450, + "dataset_size": 151455, + "size_in_bytes": 494905 + }, + "matrix_question_npi_licensor_present": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "matrix_question_npi_licensor_present", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 154657, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/matrix_question_npi_licensor_present.jsonl": { + "num_bytes": 457806, + "checksum": "657c9d6e6d50a9cd651881fcbc57cef2fd998611f6f53158ed05e48822fcd2b5" + } + }, + "download_size": 457806, + "dataset_size": 154657, + "size_in_bytes": 612463 + }, + "npi_present_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "npi_present_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 139860, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/npi_present_1.jsonl": { + "num_bytes": 438013, + "checksum": "9aa85dc5c2b0d85ea51f21db50d6d327ab5b45d1669e386d54b1360d3ac954af" + } + }, + "download_size": 438013, + "dataset_size": 139860, + "size_in_bytes": 577873 + }, + "npi_present_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "npi_present_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 129031, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/npi_present_2.jsonl": { + "num_bytes": 422136, + "checksum": "52fcfa230a40ae27998273cc08ecd062aad17025c92389df3404a85ed519909b" + } + }, + "download_size": 422136, + "dataset_size": 129031, + "size_in_bytes": 551167 + }, + "only_npi_licensor_present": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "only_npi_licensor_present", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 149911, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/only_npi_licensor_present.jsonl": { + "num_bytes": 459170, + "checksum": "475c6595187a32e4dcfb868d3db90394e89669a94a9312ebd60f88dad1665f1c" + } + }, + "download_size": 459170, + "dataset_size": 149911, + "size_in_bytes": 609081 + }, + "only_npi_scope": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "only_npi_scope", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 210297, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/only_npi_scope.jsonl": { + "num_bytes": 583720, + "checksum": "ca1f0671643e37936515f02125036bc2904ab9b35c4689b0e7a3ea3914e571f6" + } + }, + "download_size": 583720, + "dataset_size": 210297, + "size_in_bytes": 794017 + }, + "passive_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "passive_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 147277, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/passive_1.jsonl": { + "num_bytes": 340272, + "checksum": "e9374307bca5c24e9df18597a0859f7fb5ca4cfe5f136ec894d1ba69f4c90116" + } + }, + "download_size": 340272, + "dataset_size": 147277, + "size_in_bytes": 487549 + }, + "passive_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "passive_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 115355, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/passive_2.jsonl": { + "num_bytes": 308350, + "checksum": "4d49e438185bed4122a252dd9ee69bc8eccb2fe6c1b09ccc78198fd925fed8c5" + } + }, + "download_size": 308350, + "dataset_size": 115355, + "size_in_bytes": 423705 + }, + "principle_A_c_command": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "principle_A_c_command", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 189885, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_c_command.jsonl": { + "num_bytes": 527689, + "checksum": "74c8111666542c797dee1c38c9e432b0b65751dfa0faa51874a990f2cbfc76fc" + } + }, + "download_size": 527689, + "dataset_size": 189885, + "size_in_bytes": 717574 + }, + "principle_A_case_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "principle_A_case_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 171793, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_case_1.jsonl": { + "num_bytes": 477239, + "checksum": "7efb5246fe315dc2cf56b80739783cffcddee9e52650716cf5aa0f6987f7644d" + } + }, + "download_size": 477239, + "dataset_size": 171793, + "size_in_bytes": 649032 + }, + "principle_A_case_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "principle_A_case_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 171807, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_case_2.jsonl": { + "num_bytes": 492973, + "checksum": "5de6db4c21a8433bc1f1030e81d5ee151efce02078460f7de7ca5e9566afd914" + } + }, + "download_size": 492973, + "dataset_size": 171807, + "size_in_bytes": 664780 + }, + "principle_A_domain_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "principle_A_domain_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 172565, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_domain_1.jsonl": { + "num_bytes": 499865, + "checksum": "2276dc35a51396c7d11a401e378e0f4b31cd10d7916c73e9a361703cb96eddc5" + } + }, + "download_size": 499865, + "dataset_size": 172565, + "size_in_bytes": 672430 + }, + "principle_A_domain_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "principle_A_domain_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 166728, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_domain_2.jsonl": { + "num_bytes": 493189, + "checksum": "9d14e5ae911fb8ddf571484ef2d6a6e6a053765c1c6d34253aec2c11f7d7306a" + } + }, + "download_size": 493189, + "dataset_size": 166728, + "size_in_bytes": 659917 + }, + "principle_A_domain_3": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "principle_A_domain_3", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 160393, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_domain_3.jsonl": { + "num_bytes": 513886, + "checksum": "a0d66b9d8541ab1df08f629161c6c4e8bca3aa2693eaa342edd2097e6b579bdd" + } + }, + "download_size": 513886, + "dataset_size": 160393, + "size_in_bytes": 674279 + }, + "principle_A_reconstruction": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "principle_A_reconstruction", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 153499, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/principle_A_reconstruction.jsonl": { + "num_bytes": 345494, + "checksum": "65dc403c4d6e9a609f2c4cc2ec6c8baa05cf6fce89b920a816b0d9932204b2a3" + } + }, + "download_size": 345494, + "dataset_size": 153499, + "size_in_bytes": 498993 + }, + "regular_plural_subject_verb_agreement_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "regular_plural_subject_verb_agreement_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 160214, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/regular_plural_subject_verb_agreement_1.jsonl": { + "num_bytes": 451850, + "checksum": "1a18d94062c8e792c0a200a3b00dff0e051a92bb084c05e9b5e5c73bee32a620" + } + }, + "download_size": 451850, + "dataset_size": 160214, + "size_in_bytes": 612064 + }, + "regular_plural_subject_verb_agreement_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "regular_plural_subject_verb_agreement_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 155004, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/regular_plural_subject_verb_agreement_2.jsonl": { + "num_bytes": 456477, + "checksum": "e5b4668811e63b00ac4c9569cb9553281a39c21df43256905ad78e59ae24942c" + } + }, + "download_size": 456477, + "dataset_size": 155004, + "size_in_bytes": 611481 + }, + "sentential_negation_npi_licensor_present": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "sentential_negation_npi_licensor_present", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 173259, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/sentential_negation_npi_licensor_present.jsonl": { + "num_bytes": 490996, + "checksum": "c43405aca13ba4aa6ad87f9b051e4ba5fc458f24e683cfe0af23e6e09ec7268b" + } + }, + "download_size": 490996, + "dataset_size": 173259, + "size_in_bytes": 664255 + }, + "sentential_negation_npi_scope": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "sentential_negation_npi_scope", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 233493, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/sentential_negation_npi_scope.jsonl": { + "num_bytes": 614930, + "checksum": "6c341ccc39c94bea45c0412986751bf2f6eca48626728e30e23d2b6a22c8b949" + } + }, + "download_size": 614930, + "dataset_size": 233493, + "size_in_bytes": 848423 + }, + "sentential_subject_island": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "sentential_subject_island", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 173827, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/sentential_subject_island.jsonl": { + "num_bytes": 365822, + "checksum": "a286b67cc5912aafeb06750b38923abf3338786ffa4309e0a6a07b1dbe2444bd" + } + }, + "download_size": 365822, + "dataset_size": 173827, + "size_in_bytes": 539649 + }, + "superlative_quantifiers_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "superlative_quantifiers_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 160685, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/superlative_quantifiers_1.jsonl": { + "num_bytes": 381189, + "checksum": "bebc5c5ff28cf226cc35293c60cde1fcf7565a7b58cb219d01ac8908b952a497" + } + }, + "download_size": 381189, + "dataset_size": 160685, + "size_in_bytes": 541874 + }, + "superlative_quantifiers_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "superlative_quantifiers_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 160735, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/superlative_quantifiers_2.jsonl": { + "num_bytes": 516120, + "checksum": "4bef08f3e6bd3fbb46021dda34000f3c00222c5576babe4882a8541f6820a106" + } + }, + "download_size": 516120, + "dataset_size": 160735, + "size_in_bytes": 676855 + }, + "tough_vs_raising_1": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "tough_vs_raising_1", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 150031, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/tough_vs_raising_1.jsonl": { + "num_bytes": 343026, + "checksum": "8eb3a1ecbdbd969c5b77e2f85b12a39d33723bdc58d673e5067bb456da91d0e5" + } + }, + "download_size": 343026, + "dataset_size": 150031, + "size_in_bytes": 493057 + }, + "tough_vs_raising_2": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "tough_vs_raising_2", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 171079, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/tough_vs_raising_2.jsonl": { + "num_bytes": 364074, + "checksum": "a6db2f24a8aee0c90a9be63e6bb6787fe04e990c518761115f69f27c7438b5ae" + } + }, + "download_size": 364074, + "dataset_size": 171079, + "size_in_bytes": 535153 + }, + "transitive": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "transitive", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 134499, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/transitive.jsonl": { + "num_bytes": 460291, + "checksum": "48de2025571897c7590659c793d2d1ce85c247808d4199bad60f66cf2eb179ef" + } + }, + "download_size": 460291, + "dataset_size": 134499, + "size_in_bytes": 594790 + }, + "wh_island": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "wh_island", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 143735, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_island.jsonl": { + "num_bytes": 448630, + "checksum": "77824ca8c0008fb7a204e785ca2dcee1b2803f666bb545bd95cb8a51e04f1a4c" + } + }, + "download_size": 448630, + "dataset_size": 143735, + "size_in_bytes": 592365 + }, + "wh_questions_object_gap": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "wh_questions_object_gap", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 194440, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_questions_object_gap.jsonl": { + "num_bytes": 387435, + "checksum": "685c74eb8c199f30a24bba1bfabf9d9dff7962a64430e62b9b5f61d5369241ff" + } + }, + "download_size": 387435, + "dataset_size": 194440, + "size_in_bytes": 581875 + }, + "wh_questions_subject_gap": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "wh_questions_subject_gap", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 196988, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_questions_subject_gap.jsonl": { + "num_bytes": 389983, + "checksum": "3c33e4efc7182824211ff3288694cf010f4ce84f85c8df67410d784139c722b3" + } + }, + "download_size": 389983, + "dataset_size": 196988, + "size_in_bytes": 586971 + }, + "wh_questions_subject_gap_long_distance": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "wh_questions_subject_gap_long_distance", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 269665, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_questions_subject_gap_long_distance.jsonl": { + "num_bytes": 462660, + "checksum": "985bae202dbc5566b9b5c51bd77dd9d1dbcd87e0f38e008c34a11d7aa07db788" + } + }, + "download_size": 462660, + "dataset_size": 269665, + "size_in_bytes": 732325 + }, + "wh_vs_that_no_gap": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "wh_vs_that_no_gap", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 190267, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_vs_that_no_gap.jsonl": { + "num_bytes": 383262, + "checksum": "6a25d0d1499e9f522696a62ce9dba8c49f58c5446fddc6d852bf562d1016163c" + } + }, + "download_size": 383262, + "dataset_size": 190267, + "size_in_bytes": 573529 + }, + "wh_vs_that_no_gap_long_distance": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "wh_vs_that_no_gap_long_distance", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 248434, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_vs_that_no_gap_long_distance.jsonl": { + "num_bytes": 441429, + "checksum": "968faf59602a67f9cfd7c2e91a314a3f7342fd5daf41c21cadd25288fc85fdcf" + } + }, + "download_size": 441429, + "dataset_size": 248434, + "size_in_bytes": 689863 + }, + "wh_vs_that_with_gap": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "wh_vs_that_with_gap", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 174781, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_vs_that_with_gap.jsonl": { + "num_bytes": 367776, + "checksum": "56f1ca988a8ca368234f12359792c414fdfec2891154e3f8ce6404da1d3f3212" + } + }, + "download_size": 367776, + "dataset_size": 174781, + "size_in_bytes": 542557 + }, + "wh_vs_that_with_gap_long_distance": { + "description": "\nBLiMP is a challenge set for evaluating what language models (LMs) know about\nmajor grammatical phenomena in English. BLiMP consists of 67 sub-datasets, each\ncontaining 1000 minimal pairs isolating specific contrasts in syntax,\nmorphology, or semantics. The data is automatically generated according to\nexpert-crafted grammars.\n", + "citation": "\n@article{warstadt2019blimp,\n title={BLiMP: A Benchmark of Linguistic Minimal Pairs for English},\n author={Warstadt, Alex and Parrish, Alicia and Liu, Haokun and Mohananey, Anhad and Peng, Wei, and Wang, Sheng-Fu and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1912.00582},\n year={2019}\n}\n", + "homepage": "https://github.com/alexwarstadt/blimp/tree/master/", + "license": "", + "features": { + "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" + } + }, + "supervised_keys": null, + "builder_name": "blimp", + "config_name": "wh_vs_that_with_gap_long_distance", + "version": { + "version_str": "0.1.0", + "description": null, + "datasets_version_to_prepare": null, + "major": 0, + "minor": 1, + "patch": 0 + }, + "splits": { + "train": { + "name": "train", + "num_bytes": 232990, + "num_examples": 1000, + "dataset_name": "blimp" + } + }, + "download_checksums": { + "https://raw.githubusercontent.com/alexwarstadt/blimp/master/data/wh_vs_that_with_gap_long_distance.jsonl": { + "num_bytes": 425985, + "checksum": "3cb19100d6384264e512801e96c20e172a3c0c7778276b80ed60c65e97b86ab5" + } + }, + "download_size": 425985, + "dataset_size": 232990, + "size_in_bytes": 658975 + } +} \ No newline at end of file