Datasets:
File size: 97,991 Bytes
c8e2ca7 |
1 |
{"boolq_sv": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nBoolQ (Boolean Questions, Clark et al., 2019a) is a QA task where each example consists of a short\npassage and a yes/no question about the passage. The questions are provided anonymously and\nunsolicited by users of the Google search engine, and afterwards paired with a paragraph from a\nWikipedia article containing the answer. Following the original work, we evaluate with accuracy.", "citation": "@inproceedings{clark2019boolq,\n title={BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions},\n author={Clark, Christopher and Lee, Kenton and Chang, Ming-Wei, and Kwiatkowski, Tom and Collins, Michael, and Toutanova, Kristina},\n booktitle={NAACL},\n year={2019}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"question": {"dtype": "string", "id": null, "_type": "Value"}, "passage": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "boolq_sv", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 4211792, "num_examples": 6285, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 2057950, "num_examples": 3142, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 2150348, "num_examples": 3270, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/sv/boolq.tar.gz": {"num_bytes": 3368913, "checksum": "159555cd9b2c93ed20480ee3bd2b7e3cb016b7929b81e121fbd0f2d1a030b074"}}, "download_size": 3368913, "post_processing_size": null, "dataset_size": 8420090, "size_in_bytes": 11789003}, "cb_sv": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe CommitmentBank (De Marneffe et al., 2019) is a corpus of short texts in which at least\none sentence contains an embedded clause. Each of these embedded clauses is annotated with the\ndegree to which we expect that the person who wrote the text is committed to the truth of the clause.\nThe resulting task framed as three-class textual entailment on examples that are drawn from the Wall\nStreet Journal, fiction from the British National Corpus, and Switchboard. Each example consists\nof a premise containing an embedded clause and the corresponding hypothesis is the extraction of\nthat clause. We use a subset of the data that had inter-annotator agreement above 0.85. The data is\nimbalanced (relatively fewer neutral examples), so we evaluate using accuracy and F1, where for\nmulti-class F1 we compute the unweighted average of the F1 per class.", "citation": "@article{de marneff_simons_tonhauser_2019,\n title={The CommitmentBank: Investigating projection in naturally occurring discourse},\n journal={proceedings of Sinn und Bedeutung 23},\n author={De Marneff, Marie-Catherine and Simons, Mandy and Tonhauser, Judith},\n year={2019}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "cb_sv", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 72247, "num_examples": 201, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 18479, "num_examples": 49, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 22901, "num_examples": 56, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/sv/cb.tar.gz": {"num_bytes": 41036, "checksum": "85587d9ab0ca5cadbb1f1ebad65b648a79c65baff84989334444f503b451752e"}}, "download_size": 41036, "post_processing_size": null, "dataset_size": 113627, "size_in_bytes": 154663}, "copa_sv": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Choice Of Plausible Alternatives (COPA, Roemmele et al., 2011) dataset is a causal\nreasoning task in which a system is given a premise sentence and two possible alternatives. The\nsystem must choose the alternative which has the more plausible causal relationship with the premise.\nThe method used for the construction of the alternatives ensures that the task requires causal reasoning\nto solve. Examples either deal with alternative possible causes or alternative possible effects of the\npremise sentence, accompanied by a simple question disambiguating between the two instance\ntypes for the model. All examples are handcrafted and focus on topics from online blogs and a\nphotography-related encyclopedia. Following the recommendation of the authors, we evaluate using\naccuracy.", "citation": "@inproceedings{roemmele2011choice,\n title={Choice of plausible alternatives: An evaluation of commonsense causal reasoning},\n author={Roemmele, Melissa and Bejan, Cosmin Adrian and Gordon, Andrew S},\n booktitle={2011 AAAI Spring Symposium Series},\n year={2011}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "choice1": {"dtype": "string", "id": null, "_type": "Value"}, "choice2": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "copa_sv", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 38614, "num_examples": 321, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 9447, "num_examples": 79, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 12258, "num_examples": 100, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/sv/copa.tar.gz": {"num_bytes": 22790, "checksum": "ddea054e155a16c8857ed078073943058b4fe2ca4c59155d52afafc878ce5722"}}, "download_size": 22790, "post_processing_size": null, "dataset_size": 60319, "size_in_bytes": 83109}, "rte_sv": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Recognizing Textual Entailment (RTE) datasets come from a series of annual competitions\non textual entailment, the problem of predicting whether a given premise sentence entails a given\nhypothesis sentence (also known as natural language inference, NLI). RTE was previously included\nin GLUE, and we use the same data and format as before: We merge data from RTE1 (Dagan\net al., 2006), RTE2 (Bar Haim et al., 2006), RTE3 (Giampiccolo et al., 2007), and RTE5 (Bentivogli\net al., 2009). All datasets are combined and converted to two-class classification: entailment and\nnot_entailment. Of all the GLUE tasks, RTE was among those that benefited from transfer learning\nthe most, jumping from near random-chance performance (~56%) at the time of GLUE's launch to\n85% accuracy (Liu et al., 2019c) at the time of writing. Given the eight point gap with respect to\nhuman performance, however, the task is not yet solved by machines, and we expect the remaining\ngap to be difficult to close.", "citation": "@inproceedings{dagan2005pascal,\n title={The PASCAL recognising textual entailment challenge},\n author={Dagan, Ido and Glickman, Oren and Magnini, Bernardo},\n booktitle={Machine Learning Challenges Workshop},\n pages={177--190},\n year={2005},\n organization={Springer}\n}\n@inproceedings{bar2006second,\n title={The second pascal recognising textual entailment challenge},\n author={Bar-Haim, Roy and Dagan, Ido and Dolan, Bill and Ferro, Lisa and Giampiccolo, Danilo and Magnini, Bernardo and Szpektor, Idan},\n booktitle={Proceedings of the second PASCAL challenges workshop on recognising textual entailment},\n volume={6},\n number={1},\n pages={6--4},\n year={2006},\n organization={Venice}\n}\n@inproceedings{giampiccolo2007third,\n title={The third pascal recognizing textual entailment challenge},\n author={Giampiccolo, Danilo and Magnini, Bernardo and Dagan, Ido and Dolan, Bill},\n booktitle={Proceedings of the ACL-PASCAL workshop on textual entailment and paraphrasing},\n pages={1--9},\n year={2007},\n organization={Association for Computational Linguistics}\n}\n@inproceedings{bentivogli2009fifth,\n title={The Fifth PASCAL Recognizing Textual Entailment Challenge.},\n author={Bentivogli, Luisa and Clark, Peter and Dagan, Ido and Giampiccolo, Danilo},\n booktitle={TAC},\n year={2009}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "rte_sv", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 798331, "num_examples": 2214, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 91197, "num_examples": 276, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 95768, "num_examples": 277, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/sv/rte.tar.gz": {"num_bytes": 395529, "checksum": "e77ce27a918be490ddf3f39dd7b25ae9dca202e82bd037b01fca0dc5ddedf8d0"}}, "download_size": 395529, "post_processing_size": null, "dataset_size": 985296, "size_in_bytes": 1380825}, "qqp_sv": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Quora Question Pairs2 dataset is a collection of question pairs from the\ncommunity question-answering website Quora. The task is to determine whether a\npair of questions are semantically equivalent.", "citation": "@online{WinNT,\nauthor = {Iyer, Shankar and Dandekar, Nikhil and Csernai, Kornel},\ntitle = {First Quora Dataset Release: Question Pairs},\nyear = {2017},\nurl = {https://data.quora.com/First-Quora-Dataset-Release-Question-Pairs},\nurldate = {2019-04-03}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"text_a": {"dtype": "string", "id": null, "_type": "Value"}, "text_b": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "qqp_sv", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 46419002, "num_examples": 323419, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 5806503, "num_examples": 40427, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 5795360, "num_examples": 40430, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/sv/qqp.tar.gz": {"num_bytes": 21998351, "checksum": "c54b8b76973671f5ef1ad8d87a7f2a292851c69b36a8bdbb2c93eaa05607cefb"}}, "download_size": 21998351, "post_processing_size": null, "dataset_size": 58020865, "size_in_bytes": 80019216}, "qnli_sv": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Stanford Question Answering Dataset is a question-answering\ndataset consisting of question-paragraph pairs, where one of the sentences in the paragraph (drawn\nfrom Wikipedia) contains the answer to the corresponding question (written by an annotator). We\nconvert the task into sentence pair classification by forming a pair between each question and each\nsentence in the corresponding context, and filtering out pairs with low lexical overlap between the\nquestion and the context sentence. The task is to determine whether the context sentence contains\nthe answer to the question. This modified version of the original task removes the requirement that\nthe model select the exact answer, but also removes the simplifying assumptions that the answer\nis always present in the input and that lexical overlap is a reliable cue.", "citation": "@article{rajpurkar2016squad,\n title={Squad: 100,000+ questions for machine comprehension of text},\n author={Rajpurkar, Pranav and Zhang, Jian and Lopyrev, Konstantin and Liang, Percy},\n journal={arXiv preprint arXiv:1606.05250},\n year={2016}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "qnli_sv", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 25377994, "num_examples": 99506, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 1343820, "num_examples": 5237, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 1422786, "num_examples": 5463, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/sv/qnli.tar.gz": {"num_bytes": 11218767, "checksum": "0c9205a5f0f8c2119c3e1d67e5f621201a089a39bd9335d64b51a86a87f90ad3"}}, "download_size": 11218767, "post_processing_size": null, "dataset_size": 28144600, "size_in_bytes": 39363367}, "stsb_sv": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Semantic Textual Similarity Benchmark (Cer et al., 2017) is a collection of\nsentence pairs drawn from news headlines, video and image captions, and natural\nlanguage inference data. Each pair is human-annotated with a similarity score\nfrom 1 to 5.", "citation": "@article{cer2017semeval,\n title={Semeval-2017 task 1: Semantic textual similarity-multilingual and cross-lingual focused evaluation},\n author={Cer, Daniel and Diab, Mona and Agirre, Eneko and Lopez-Gazpio, Inigo and Specia, Lucia},\n journal={arXiv preprint arXiv:1708.00055},\n year={2017}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"text_a": {"dtype": "string", "id": null, "_type": "Value"}, "text_b": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "stsb_sv", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 640930, "num_examples": 4312, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 217613, "num_examples": 1437, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 241574, "num_examples": 1500, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/sv/stsb.tar.gz": {"num_bytes": 371403, "checksum": "c8d0df38abf3089fc2a83310bd536b69dbf09375f3df90fbb513b02b549f75e0"}}, "download_size": 371403, "post_processing_size": null, "dataset_size": 1100117, "size_in_bytes": 1471520}, "mnli_sv": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Multi-Genre Natural Language Inference Corpus is a crowdsourced\ncollection of sentence pairs with textual entailment annotations. Given a premise sentence\nand a hypothesis sentence, the task is to predict whether the premise entails the hypothesis\n(entailment), contradicts the hypothesis (contradiction), or neither (neutral). The premise sentences are\ngathered from ten different sources, including transcribed speech, fiction, and government reports.\nWe use the standard test set, for which we obtained private labels from the authors, and evaluate\non both the matched (in-domain) and mismatched (cross-domain) section. We also use and recommend\nthe SNLI corpus as 550k examples of auxiliary training data.", "citation": " @InProceedings{N18-1101,\n author = \"Williams, Adina\n and Nangia, Nikita\n and Bowman, Samuel\",\n title = \"A Broad-Coverage Challenge Corpus for\n Sentence Understanding through Inference\",\n booktitle = \"Proceedings of the 2018 Conference of\n the North American Chapter of the\n Association for Computational Linguistics:\n Human Language Technologies, Volume 1 (Long\n Papers)\",\n year = \"2018\",\n publisher = \"Association for Computational Linguistics\",\n pages = \"1112--1122\",\n location = \"New Orleans, Louisiana\",\n url = \"http://aclweb.org/anthology/N18-1101\"\n }\n @article{bowman2015large,\n title={A large annotated corpus for learning natural language inference},\n author={Bowman, Samuel R and Angeli, Gabor and Potts, Christopher and Manning, Christopher D},\n journal={arXiv preprint arXiv:1508.05326},\n year={2015}\n }\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "mnli_sv", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 76720827, "num_examples": 383124, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 1906810, "num_examples": 9578, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 1941882, "num_examples": 9815, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/sv/mnli.tar.gz": {"num_bytes": 30839949, "checksum": "b5f40c836e15fdd2b2aaad3ea8792a7a8b1cb4ac6fb5baacc6cf13933cdc7319"}}, "download_size": 30839949, "post_processing_size": null, "dataset_size": 80569519, "size_in_bytes": 111409468}, "mrpc_sv": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Microsoft Research Paraphrase Corpus (Dolan & Brockett, 2005) is a corpus of\nsentence pairs automatically extracted from online news sources, with human annotations\nfor whether the sentences in the pair are semantically equivalent.", "citation": "@inproceedings{dolan2005automatically,\n title={Automatically constructing a corpus of sentential paraphrases},\n author={Dolan, William B and Brockett, Chris},\n booktitle={Proceedings of the Third International Workshop on Paraphrasing (IWP2005)},\n year={2005}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"text_a": {"dtype": "string", "id": null, "_type": "Value"}, "text_b": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "mrpc_sv", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 853395, "num_examples": 3261, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 105267, "num_examples": 407, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 108361, "num_examples": 408, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/sv/mrpc.tar.gz": {"num_bytes": 381946, "checksum": "13900c98246c7389bf8eda4d54e82042a6c1aed095ad26b60582919b7e8380a4"}}, "download_size": 381946, "post_processing_size": null, "dataset_size": 1067023, "size_in_bytes": 1448969}, "wnli_sv": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Winograd Schema Challenge (Levesque et al., 2011) is a reading comprehension task\nin which a system must read a sentence with a pronoun and select the referent of that pronoun from\na list of choices. The examples are manually constructed to foil simple statistical methods: Each\none is contingent on contextual information provided by a single word or phrase in the sentence.\nTo convert the problem into sentence pair classification, we construct sentence pairs by replacing\nthe ambiguous pronoun with each possible referent. The task is to predict if the sentence with the\npronoun substituted is entailed by the original sentence. We use a small evaluation set consisting of\nnew examples derived from fiction books that was shared privately by the authors of the original\ncorpus. While the included training set is balanced between two classes, the test set is imbalanced\nbetween them (65% not entailment). Also, due to a data quirk, the development set is adversarial:\nhypotheses are sometimes shared between training and development examples, so if a model memorizes the\ntraining examples, they will predict the wrong label on corresponding development set\nexample. As with QNLI, each example is evaluated separately, so there is not a systematic correspondence\nbetween a model's score on this task and its score on the unconverted original task. We\ncall converted dataset WNLI (Winograd NLI).", "citation": "@inproceedings{levesque2012winograd,\n title={The winograd schema challenge},\n author={Levesque, Hector and Davis, Ernest and Morgenstern, Leora},\n booktitle={Thirteenth International Conference on the Principles of Knowledge Representation and Reasoning},\n year={2012}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "wnli_sv", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 94959, "num_examples": 565, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 11665, "num_examples": 70, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 12020, "num_examples": 71, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/sv/wnli.tar.gz": {"num_bytes": 30427, "checksum": "2e4b019e3b2f614a1ef12a0b243d3d1b08cf2d9bf79676998ff259447a3461de"}}, "download_size": 30427, "post_processing_size": null, "dataset_size": 118644, "size_in_bytes": 149071}, "sst_sv": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Stanford Sentiment Treebank consists of sentences from movie reviews and\nhuman annotations of their sentiment. The task is to predict the sentiment of a\ngiven sentence. We use the two-way (positive/negative) class split, and use only\nsentence-level labels.", "citation": "@inproceedings{socher2013recursive,\n title={Recursive deep models for semantic compositionality over a sentiment treebank},\n author={Socher, Richard and Perelygin, Alex and Wu, Jean and Chuang, Jason and Manning, Christopher D and Ng, Andrew and Potts, Christopher},\n booktitle={Proceedings of the 2013 conference on empirical methods in natural language processing},\n pages={1631--1642},\n year={2013}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "sst_sv", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 4526277, "num_examples": 66486, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 58238, "num_examples": 863, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 105918, "num_examples": 872, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/sv/sst.tar.gz": {"num_bytes": 1984447, "checksum": "86ea38b24a963e42a07ee88e822efa262c674580d3e8f368a4b4179761ca0d81"}}, "download_size": 1984447, "post_processing_size": null, "dataset_size": 4690433, "size_in_bytes": 6674880}, "boolq_nb": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nBoolQ (Boolean Questions, Clark et al., 2019a) is a QA task where each example consists of a short\npassage and a yes/no question about the passage. The questions are provided anonymously and\nunsolicited by users of the Google search engine, and afterwards paired with a paragraph from a\nWikipedia article containing the answer. Following the original work, we evaluate with accuracy.", "citation": "@inproceedings{clark2019boolq,\n title={BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions},\n author={Clark, Christopher and Lee, Kenton and Chang, Ming-Wei, and Kwiatkowski, Tom and Collins, Michael, and Toutanova, Kristina},\n booktitle={NAACL},\n year={2019}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"question": {"dtype": "string", "id": null, "_type": "Value"}, "passage": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "boolq_nb", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 3966994, "num_examples": 6285, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 1938942, "num_examples": 3142, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 2024171, "num_examples": 3270, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/nb/boolq.tar.gz": {"num_bytes": 3254901, "checksum": "0038ecb49187122ecd3e4607ca3a2db2afae27fd8293279ec805d809d5567de8"}}, "download_size": 3254901, "post_processing_size": null, "dataset_size": 7930107, "size_in_bytes": 11185008}, "cb_nb": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe CommitmentBank (De Marneffe et al., 2019) is a corpus of short texts in which at least\none sentence contains an embedded clause. Each of these embedded clauses is annotated with the\ndegree to which we expect that the person who wrote the text is committed to the truth of the clause.\nThe resulting task framed as three-class textual entailment on examples that are drawn from the Wall\nStreet Journal, fiction from the British National Corpus, and Switchboard. Each example consists\nof a premise containing an embedded clause and the corresponding hypothesis is the extraction of\nthat clause. We use a subset of the data that had inter-annotator agreement above 0.85. The data is\nimbalanced (relatively fewer neutral examples), so we evaluate using accuracy and F1, where for\nmulti-class F1 we compute the unweighted average of the F1 per class.", "citation": "@article{de marneff_simons_tonhauser_2019,\n title={The CommitmentBank: Investigating projection in naturally occurring discourse},\n journal={proceedings of Sinn und Bedeutung 23},\n author={De Marneff, Marie-Catherine and Simons, Mandy and Tonhauser, Judith},\n year={2019}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "cb_nb", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 68229, "num_examples": 201, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 17578, "num_examples": 49, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 21560, "num_examples": 56, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/nb/cb.tar.gz": {"num_bytes": 39777, "checksum": "7d4a11bbd79bdfafad0b7a4c0f5d4f8cb4c1d944f6bac8dfe6782bff7725a5da"}}, "download_size": 39777, "post_processing_size": null, "dataset_size": 107367, "size_in_bytes": 147144}, "copa_nb": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Choice Of Plausible Alternatives (COPA, Roemmele et al., 2011) dataset is a causal\nreasoning task in which a system is given a premise sentence and two possible alternatives. The\nsystem must choose the alternative which has the more plausible causal relationship with the premise.\nThe method used for the construction of the alternatives ensures that the task requires causal reasoning\nto solve. Examples either deal with alternative possible causes or alternative possible effects of the\npremise sentence, accompanied by a simple question disambiguating between the two instance\ntypes for the model. All examples are handcrafted and focus on topics from online blogs and a\nphotography-related encyclopedia. Following the recommendation of the authors, we evaluate using\naccuracy.", "citation": "@inproceedings{roemmele2011choice,\n title={Choice of plausible alternatives: An evaluation of commonsense causal reasoning},\n author={Roemmele, Melissa and Bejan, Cosmin Adrian and Gordon, Andrew S},\n booktitle={2011 AAAI Spring Symposium Series},\n year={2011}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "choice1": {"dtype": "string", "id": null, "_type": "Value"}, "choice2": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "copa_nb", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 37790, "num_examples": 321, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 9173, "num_examples": 79, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 12044, "num_examples": 100, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/nb/copa.tar.gz": {"num_bytes": 22695, "checksum": "743dc02ade29e7cb4cb77f6716fcfd949635114f889ff3daec6f09bb6592f541"}}, "download_size": 22695, "post_processing_size": null, "dataset_size": 59007, "size_in_bytes": 81702}, "rte_nb": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Recognizing Textual Entailment (RTE) datasets come from a series of annual competitions\non textual entailment, the problem of predicting whether a given premise sentence entails a given\nhypothesis sentence (also known as natural language inference, NLI). RTE was previously included\nin GLUE, and we use the same data and format as before: We merge data from RTE1 (Dagan\net al., 2006), RTE2 (Bar Haim et al., 2006), RTE3 (Giampiccolo et al., 2007), and RTE5 (Bentivogli\net al., 2009). All datasets are combined and converted to two-class classification: entailment and\nnot_entailment. Of all the GLUE tasks, RTE was among those that benefited from transfer learning\nthe most, jumping from near random-chance performance (~56%) at the time of GLUE's launch to\n85% accuracy (Liu et al., 2019c) at the time of writing. Given the eight point gap with respect to\nhuman performance, however, the task is not yet solved by machines, and we expect the remaining\ngap to be difficult to close.", "citation": "@inproceedings{dagan2005pascal,\n title={The PASCAL recognising textual entailment challenge},\n author={Dagan, Ido and Glickman, Oren and Magnini, Bernardo},\n booktitle={Machine Learning Challenges Workshop},\n pages={177--190},\n year={2005},\n organization={Springer}\n}\n@inproceedings{bar2006second,\n title={The second pascal recognising textual entailment challenge},\n author={Bar-Haim, Roy and Dagan, Ido and Dolan, Bill and Ferro, Lisa and Giampiccolo, Danilo and Magnini, Bernardo and Szpektor, Idan},\n booktitle={Proceedings of the second PASCAL challenges workshop on recognising textual entailment},\n volume={6},\n number={1},\n pages={6--4},\n year={2006},\n organization={Venice}\n}\n@inproceedings{giampiccolo2007third,\n title={The third pascal recognizing textual entailment challenge},\n author={Giampiccolo, Danilo and Magnini, Bernardo and Dagan, Ido and Dolan, Bill},\n booktitle={Proceedings of the ACL-PASCAL workshop on textual entailment and paraphrasing},\n pages={1--9},\n year={2007},\n organization={Association for Computational Linguistics}\n}\n@inproceedings{bentivogli2009fifth,\n title={The Fifth PASCAL Recognizing Textual Entailment Challenge.},\n author={Bentivogli, Luisa and Clark, Peter and Dagan, Ido and Giampiccolo, Danilo},\n booktitle={TAC},\n year={2009}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "rte_nb", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 745583, "num_examples": 2214, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 85478, "num_examples": 276, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 89644, "num_examples": 277, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/nb/rte.tar.gz": {"num_bytes": 379837, "checksum": "c764579d7ac464fae4b295894a9e36c188707283ae02b8e0dfdcf9e86caf84ab"}}, "download_size": 379837, "post_processing_size": null, "dataset_size": 920705, "size_in_bytes": 1300542}, "qqp_nb": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Quora Question Pairs2 dataset is a collection of question pairs from the\ncommunity question-answering website Quora. The task is to determine whether a\npair of questions are semantically equivalent.", "citation": "@online{WinNT,\nauthor = {Iyer, Shankar and Dandekar, Nikhil and Csernai, Kornel},\ntitle = {First Quora Dataset Release: Question Pairs},\nyear = {2017},\nurl = {https://data.quora.com/First-Quora-Dataset-Release-Question-Pairs},\nurldate = {2019-04-03}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"text_a": {"dtype": "string", "id": null, "_type": "Value"}, "text_b": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "qqp_nb", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 44922311, "num_examples": 323419, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 5616459, "num_examples": 40427, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 5608850, "num_examples": 40430, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/nb/qqp.tar.gz": {"num_bytes": 21097603, "checksum": "f2b0000dfb7f68a277b5b961fe40d78ede324d4fd455a868919e5795cfb41d11"}}, "download_size": 21097603, "post_processing_size": null, "dataset_size": 56147620, "size_in_bytes": 77245223}, "qnli_nb": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Stanford Question Answering Dataset is a question-answering\ndataset consisting of question-paragraph pairs, where one of the sentences in the paragraph (drawn\nfrom Wikipedia) contains the answer to the corresponding question (written by an annotator). We\nconvert the task into sentence pair classification by forming a pair between each question and each\nsentence in the corresponding context, and filtering out pairs with low lexical overlap between the\nquestion and the context sentence. The task is to determine whether the context sentence contains\nthe answer to the question. This modified version of the original task removes the requirement that\nthe model select the exact answer, but also removes the simplifying assumptions that the answer\nis always present in the input and that lexical overlap is a reliable cue.", "citation": "@article{rajpurkar2016squad,\n title={Squad: 100,000+ questions for machine comprehension of text},\n author={Rajpurkar, Pranav and Zhang, Jian and Lopyrev, Konstantin and Liang, Percy},\n journal={arXiv preprint arXiv:1606.05250},\n year={2016}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "qnli_nb", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 24131580, "num_examples": 99506, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 1280654, "num_examples": 5237, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 1353617, "num_examples": 5463, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/nb/qnli.tar.gz": {"num_bytes": 10724704, "checksum": "604c6ccc69b081c3ff24072d97918b547e58f4b3c4b744072c1b750068930088"}}, "download_size": 10724704, "post_processing_size": null, "dataset_size": 26765851, "size_in_bytes": 37490555}, "stsb_nb": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Semantic Textual Similarity Benchmark (Cer et al., 2017) is a collection of\nsentence pairs drawn from news headlines, video and image captions, and natural\nlanguage inference data. Each pair is human-annotated with a similarity score\nfrom 1 to 5.", "citation": "@article{cer2017semeval,\n title={Semeval-2017 task 1: Semantic textual similarity-multilingual and cross-lingual focused evaluation},\n author={Cer, Daniel and Diab, Mona and Agirre, Eneko and Lopez-Gazpio, Inigo and Specia, Lucia},\n journal={arXiv preprint arXiv:1708.00055},\n year={2017}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"text_a": {"dtype": "string", "id": null, "_type": "Value"}, "text_b": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "stsb_nb", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 604413, "num_examples": 4312, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 204206, "num_examples": 1437, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 229571, "num_examples": 1500, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/nb/stsb.tar.gz": {"num_bytes": 368825, "checksum": "ddb82ab7453e90010bd27fe470a887a02e96305907994bfce39230d9379b00b8"}}, "download_size": 368825, "post_processing_size": null, "dataset_size": 1038190, "size_in_bytes": 1407015}, "mnli_nb": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Multi-Genre Natural Language Inference Corpus is a crowdsourced\ncollection of sentence pairs with textual entailment annotations. Given a premise sentence\nand a hypothesis sentence, the task is to predict whether the premise entails the hypothesis\n(entailment), contradicts the hypothesis (contradiction), or neither (neutral). The premise sentences are\ngathered from ten different sources, including transcribed speech, fiction, and government reports.\nWe use the standard test set, for which we obtained private labels from the authors, and evaluate\non both the matched (in-domain) and mismatched (cross-domain) section. We also use and recommend\nthe SNLI corpus as 550k examples of auxiliary training data.", "citation": " @InProceedings{N18-1101,\n author = \"Williams, Adina\n and Nangia, Nikita\n and Bowman, Samuel\",\n title = \"A Broad-Coverage Challenge Corpus for\n Sentence Understanding through Inference\",\n booktitle = \"Proceedings of the 2018 Conference of\n the North American Chapter of the\n Association for Computational Linguistics:\n Human Language Technologies, Volume 1 (Long\n Papers)\",\n year = \"2018\",\n publisher = \"Association for Computational Linguistics\",\n pages = \"1112--1122\",\n location = \"New Orleans, Louisiana\",\n url = \"http://aclweb.org/anthology/N18-1101\"\n }\n @article{bowman2015large,\n title={A large annotated corpus for learning natural language inference},\n author={Bowman, Samuel R and Angeli, Gabor and Potts, Christopher and Manning, Christopher D},\n journal={arXiv preprint arXiv:1508.05326},\n year={2015}\n }\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "mnli_nb", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 72938596, "num_examples": 383124, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 1820060, "num_examples": 9578, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 1842160, "num_examples": 9815, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/nb/mnli.tar.gz": {"num_bytes": 29771448, "checksum": "7bdcd02d1cb2c5fd3996011032aa1a4eef96a4dae6b4d812f96ebd0a5fcd1349"}}, "download_size": 29771448, "post_processing_size": null, "dataset_size": 76600816, "size_in_bytes": 106372264}, "mrpc_nb": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Microsoft Research Paraphrase Corpus (Dolan & Brockett, 2005) is a corpus of\nsentence pairs automatically extracted from online news sources, with human annotations\nfor whether the sentences in the pair are semantically equivalent.", "citation": "@inproceedings{dolan2005automatically,\n title={Automatically constructing a corpus of sentential paraphrases},\n author={Dolan, William B and Brockett, Chris},\n booktitle={Proceedings of the Third International Workshop on Paraphrasing (IWP2005)},\n year={2005}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"text_a": {"dtype": "string", "id": null, "_type": "Value"}, "text_b": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "mrpc_nb", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 784715, "num_examples": 3261, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 97040, "num_examples": 407, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 98674, "num_examples": 408, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/nb/mrpc.tar.gz": {"num_bytes": 368694, "checksum": "3af04ec9abc76e422562751c44edbf6bea4341ed64b231e11d9b05c414d81a15"}}, "download_size": 368694, "post_processing_size": null, "dataset_size": 980429, "size_in_bytes": 1349123}, "wnli_nb": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Winograd Schema Challenge (Levesque et al., 2011) is a reading comprehension task\nin which a system must read a sentence with a pronoun and select the referent of that pronoun from\na list of choices. The examples are manually constructed to foil simple statistical methods: Each\none is contingent on contextual information provided by a single word or phrase in the sentence.\nTo convert the problem into sentence pair classification, we construct sentence pairs by replacing\nthe ambiguous pronoun with each possible referent. The task is to predict if the sentence with the\npronoun substituted is entailed by the original sentence. We use a small evaluation set consisting of\nnew examples derived from fiction books that was shared privately by the authors of the original\ncorpus. While the included training set is balanced between two classes, the test set is imbalanced\nbetween them (65% not entailment). Also, due to a data quirk, the development set is adversarial:\nhypotheses are sometimes shared between training and development examples, so if a model memorizes the\ntraining examples, they will predict the wrong label on corresponding development set\nexample. As with QNLI, each example is evaluated separately, so there is not a systematic correspondence\nbetween a model's score on this task and its score on the unconverted original task. We\ncall converted dataset WNLI (Winograd NLI).", "citation": "@inproceedings{levesque2012winograd,\n title={The winograd schema challenge},\n author={Levesque, Hector and Davis, Ernest and Morgenstern, Leora},\n booktitle={Thirteenth International Conference on the Principles of Knowledge Representation and Reasoning},\n year={2012}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "wnli_nb", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 92154, "num_examples": 565, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 11340, "num_examples": 70, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 11646, "num_examples": 71, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/nb/wnli.tar.gz": {"num_bytes": 29677, "checksum": "07f94f2f11b502ebc188f05ecb83a4f89823d5bb4d0ae4b5b10c3052ecb8fdf9"}}, "download_size": 29677, "post_processing_size": null, "dataset_size": 115140, "size_in_bytes": 144817}, "sst_nb": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Stanford Sentiment Treebank consists of sentences from movie reviews and\nhuman annotations of their sentiment. The task is to predict the sentiment of a\ngiven sentence. We use the two-way (positive/negative) class split, and use only\nsentence-level labels.", "citation": "@inproceedings{socher2013recursive,\n title={Recursive deep models for semantic compositionality over a sentiment treebank},\n author={Socher, Richard and Perelygin, Alex and Wu, Jean and Chuang, Jason and Manning, Christopher D and Ng, Andrew and Potts, Christopher},\n booktitle={Proceedings of the 2013 conference on empirical methods in natural language processing},\n pages={1631--1642},\n year={2013}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "sst_nb", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 4328256, "num_examples": 66486, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 56231, "num_examples": 863, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 99345, "num_examples": 872, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/nb/sst.tar.gz": {"num_bytes": 1905948, "checksum": "4f932332d705d01675fe565cf356ce4a22eaa2ebf4f4ebf68c3471fd43548d9c"}}, "download_size": 1905948, "post_processing_size": null, "dataset_size": 4483832, "size_in_bytes": 6389780}, "boolq_da": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nBoolQ (Boolean Questions, Clark et al., 2019a) is a QA task where each example consists of a short\npassage and a yes/no question about the passage. The questions are provided anonymously and\nunsolicited by users of the Google search engine, and afterwards paired with a paragraph from a\nWikipedia article containing the answer. Following the original work, we evaluate with accuracy.", "citation": "@inproceedings{clark2019boolq,\n title={BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions},\n author={Clark, Christopher and Lee, Kenton and Chang, Ming-Wei, and Kwiatkowski, Tom and Collins, Michael, and Toutanova, Kristina},\n booktitle={NAACL},\n year={2019}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"question": {"dtype": "string", "id": null, "_type": "Value"}, "passage": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "boolq_da", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 4147484, "num_examples": 6285, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 2024936, "num_examples": 3142, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 2114420, "num_examples": 3270, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/da/boolq.tar.gz": {"num_bytes": 3350831, "checksum": "42f1615b4e1580845e599aed7349b835f8772a0472dfa729f896f12ce0574e55"}}, "download_size": 3350831, "post_processing_size": null, "dataset_size": 8286840, "size_in_bytes": 11637671}, "cb_da": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe CommitmentBank (De Marneffe et al., 2019) is a corpus of short texts in which at least\none sentence contains an embedded clause. Each of these embedded clauses is annotated with the\ndegree to which we expect that the person who wrote the text is committed to the truth of the clause.\nThe resulting task framed as three-class textual entailment on examples that are drawn from the Wall\nStreet Journal, fiction from the British National Corpus, and Switchboard. Each example consists\nof a premise containing an embedded clause and the corresponding hypothesis is the extraction of\nthat clause. We use a subset of the data that had inter-annotator agreement above 0.85. The data is\nimbalanced (relatively fewer neutral examples), so we evaluate using accuracy and F1, where for\nmulti-class F1 we compute the unweighted average of the F1 per class.", "citation": "@article{de marneff_simons_tonhauser_2019,\n title={The CommitmentBank: Investigating projection in naturally occurring discourse},\n journal={proceedings of Sinn und Bedeutung 23},\n author={De Marneff, Marie-Catherine and Simons, Mandy and Tonhauser, Judith},\n year={2019}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "cb_da", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 70886, "num_examples": 201, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 17924, "num_examples": 49, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 22180, "num_examples": 56, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/da/cb.tar.gz": {"num_bytes": 40604, "checksum": "ff530ff490bbf68db950064add852ccc439a16544ec22780e1818bd33b365b1a"}}, "download_size": 40604, "post_processing_size": null, "dataset_size": 110990, "size_in_bytes": 151594}, "copa_da": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Choice Of Plausible Alternatives (COPA, Roemmele et al., 2011) dataset is a causal\nreasoning task in which a system is given a premise sentence and two possible alternatives. The\nsystem must choose the alternative which has the more plausible causal relationship with the premise.\nThe method used for the construction of the alternatives ensures that the task requires causal reasoning\nto solve. Examples either deal with alternative possible causes or alternative possible effects of the\npremise sentence, accompanied by a simple question disambiguating between the two instance\ntypes for the model. All examples are handcrafted and focus on topics from online blogs and a\nphotography-related encyclopedia. Following the recommendation of the authors, we evaluate using\naccuracy.", "citation": "@inproceedings{roemmele2011choice,\n title={Choice of plausible alternatives: An evaluation of commonsense causal reasoning},\n author={Roemmele, Melissa and Bejan, Cosmin Adrian and Gordon, Andrew S},\n booktitle={2011 AAAI Spring Symposium Series},\n year={2011}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "choice1": {"dtype": "string", "id": null, "_type": "Value"}, "choice2": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "copa_da", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 38625, "num_examples": 321, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 9386, "num_examples": 79, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 12189, "num_examples": 100, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/da/copa.tar.gz": {"num_bytes": 22828, "checksum": "088fb530f36b87768cd02181ab5229bf5d1e39894e54a11afea6f41998e3d0c6"}}, "download_size": 22828, "post_processing_size": null, "dataset_size": 60200, "size_in_bytes": 83028}, "rte_da": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Recognizing Textual Entailment (RTE) datasets come from a series of annual competitions\non textual entailment, the problem of predicting whether a given premise sentence entails a given\nhypothesis sentence (also known as natural language inference, NLI). RTE was previously included\nin GLUE, and we use the same data and format as before: We merge data from RTE1 (Dagan\net al., 2006), RTE2 (Bar Haim et al., 2006), RTE3 (Giampiccolo et al., 2007), and RTE5 (Bentivogli\net al., 2009). All datasets are combined and converted to two-class classification: entailment and\nnot_entailment. Of all the GLUE tasks, RTE was among those that benefited from transfer learning\nthe most, jumping from near random-chance performance (~56%) at the time of GLUE's launch to\n85% accuracy (Liu et al., 2019c) at the time of writing. Given the eight point gap with respect to\nhuman performance, however, the task is not yet solved by machines, and we expect the remaining\ngap to be difficult to close.", "citation": "@inproceedings{dagan2005pascal,\n title={The PASCAL recognising textual entailment challenge},\n author={Dagan, Ido and Glickman, Oren and Magnini, Bernardo},\n booktitle={Machine Learning Challenges Workshop},\n pages={177--190},\n year={2005},\n organization={Springer}\n}\n@inproceedings{bar2006second,\n title={The second pascal recognising textual entailment challenge},\n author={Bar-Haim, Roy and Dagan, Ido and Dolan, Bill and Ferro, Lisa and Giampiccolo, Danilo and Magnini, Bernardo and Szpektor, Idan},\n booktitle={Proceedings of the second PASCAL challenges workshop on recognising textual entailment},\n volume={6},\n number={1},\n pages={6--4},\n year={2006},\n organization={Venice}\n}\n@inproceedings{giampiccolo2007third,\n title={The third pascal recognizing textual entailment challenge},\n author={Giampiccolo, Danilo and Magnini, Bernardo and Dagan, Ido and Dolan, Bill},\n booktitle={Proceedings of the ACL-PASCAL workshop on textual entailment and paraphrasing},\n pages={1--9},\n year={2007},\n organization={Association for Computational Linguistics}\n}\n@inproceedings{bentivogli2009fifth,\n title={The Fifth PASCAL Recognizing Textual Entailment Challenge.},\n author={Bentivogli, Luisa and Clark, Peter and Dagan, Ido and Giampiccolo, Danilo},\n booktitle={TAC},\n year={2009}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "rte_da", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 789786, "num_examples": 2214, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 90099, "num_examples": 276, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 94218, "num_examples": 277, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/da/rte.tar.gz": {"num_bytes": 392939, "checksum": "e6018ee04334b57ca50fd13d4cc73c9cece96a332d368f464badbc9c374dc01e"}}, "download_size": 392939, "post_processing_size": null, "dataset_size": 974103, "size_in_bytes": 1367042}, "qqp_da": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Quora Question Pairs2 dataset is a collection of question pairs from the\ncommunity question-answering website Quora. The task is to determine whether a\npair of questions are semantically equivalent.", "citation": "@online{WinNT,\nauthor = {Iyer, Shankar and Dandekar, Nikhil and Csernai, Kornel},\ntitle = {First Quora Dataset Release: Question Pairs},\nyear = {2017},\nurl = {https://data.quora.com/First-Quora-Dataset-Release-Question-Pairs},\nurldate = {2019-04-03}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"text_a": {"dtype": "string", "id": null, "_type": "Value"}, "text_b": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "qqp_da", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 46213247, "num_examples": 323419, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 5779399, "num_examples": 40427, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 5770158, "num_examples": 40430, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/da/qqp.tar.gz": {"num_bytes": 21500446, "checksum": "dd6169eab55cdd5d921328207a5a7facd2789ecbad3dd247c6b33d981b849319"}}, "download_size": 21500446, "post_processing_size": null, "dataset_size": 57762804, "size_in_bytes": 79263250}, "qnli_da": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Stanford Question Answering Dataset is a question-answering\ndataset consisting of question-paragraph pairs, where one of the sentences in the paragraph (drawn\nfrom Wikipedia) contains the answer to the corresponding question (written by an annotator). We\nconvert the task into sentence pair classification by forming a pair between each question and each\nsentence in the corresponding context, and filtering out pairs with low lexical overlap between the\nquestion and the context sentence. The task is to determine whether the context sentence contains\nthe answer to the question. This modified version of the original task removes the requirement that\nthe model select the exact answer, but also removes the simplifying assumptions that the answer\nis always present in the input and that lexical overlap is a reliable cue.", "citation": "@article{rajpurkar2016squad,\n title={Squad: 100,000+ questions for machine comprehension of text},\n author={Rajpurkar, Pranav and Zhang, Jian and Lopyrev, Konstantin and Liang, Percy},\n journal={arXiv preprint arXiv:1606.05250},\n year={2016}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "qnli_da", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 25199883, "num_examples": 99506, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 1335353, "num_examples": 5237, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 1412721, "num_examples": 5463, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/da/qnli.tar.gz": {"num_bytes": 11157410, "checksum": "260153f22ba714106852d1545906e77e7ce490e1220a2ac286730959e37b9278"}}, "download_size": 11157410, "post_processing_size": null, "dataset_size": 27947957, "size_in_bytes": 39105367}, "stsb_da": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Semantic Textual Similarity Benchmark (Cer et al., 2017) is a collection of\nsentence pairs drawn from news headlines, video and image captions, and natural\nlanguage inference data. Each pair is human-annotated with a similarity score\nfrom 1 to 5.", "citation": "@article{cer2017semeval,\n title={Semeval-2017 task 1: Semantic textual similarity-multilingual and cross-lingual focused evaluation},\n author={Cer, Daniel and Diab, Mona and Agirre, Eneko and Lopez-Gazpio, Inigo and Specia, Lucia},\n journal={arXiv preprint arXiv:1708.00055},\n year={2017}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"text_a": {"dtype": "string", "id": null, "_type": "Value"}, "text_b": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "stsb_da", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 633017, "num_examples": 4312, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 214058, "num_examples": 1437, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 238054, "num_examples": 1500, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/da/stsb.tar.gz": {"num_bytes": 368211, "checksum": "c80b4d35b817d427f3bb2174039c44881957ad2b020e60dff6968e717fc8acdb"}}, "download_size": 368211, "post_processing_size": null, "dataset_size": 1085129, "size_in_bytes": 1453340}, "mnli_da": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Multi-Genre Natural Language Inference Corpus is a crowdsourced\ncollection of sentence pairs with textual entailment annotations. Given a premise sentence\nand a hypothesis sentence, the task is to predict whether the premise entails the hypothesis\n(entailment), contradicts the hypothesis (contradiction), or neither (neutral). The premise sentences are\ngathered from ten different sources, including transcribed speech, fiction, and government reports.\nWe use the standard test set, for which we obtained private labels from the authors, and evaluate\non both the matched (in-domain) and mismatched (cross-domain) section. We also use and recommend\nthe SNLI corpus as 550k examples of auxiliary training data.", "citation": " @InProceedings{N18-1101,\n author = \"Williams, Adina\n and Nangia, Nikita\n and Bowman, Samuel\",\n title = \"A Broad-Coverage Challenge Corpus for\n Sentence Understanding through Inference\",\n booktitle = \"Proceedings of the 2018 Conference of\n the North American Chapter of the\n Association for Computational Linguistics:\n Human Language Technologies, Volume 1 (Long\n Papers)\",\n year = \"2018\",\n publisher = \"Association for Computational Linguistics\",\n pages = \"1112--1122\",\n location = \"New Orleans, Louisiana\",\n url = \"http://aclweb.org/anthology/N18-1101\"\n }\n @article{bowman2015large,\n title={A large annotated corpus for learning natural language inference},\n author={Bowman, Samuel R and Angeli, Gabor and Potts, Christopher and Manning, Christopher D},\n journal={arXiv preprint arXiv:1508.05326},\n year={2015}\n }\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "mnli_da", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 76027939, "num_examples": 383124, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 1890232, "num_examples": 9578, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 1919687, "num_examples": 9815, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/da/mnli.tar.gz": {"num_bytes": 30837170, "checksum": "f96fd8d1b027c56f04ae1b21eb53415b39ad6ee0f97b2e2225d82794d3be350d"}}, "download_size": 30837170, "post_processing_size": null, "dataset_size": 79837858, "size_in_bytes": 110675028}, "mrpc_da": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Microsoft Research Paraphrase Corpus (Dolan & Brockett, 2005) is a corpus of\nsentence pairs automatically extracted from online news sources, with human annotations\nfor whether the sentences in the pair are semantically equivalent.", "citation": "@inproceedings{dolan2005automatically,\n title={Automatically constructing a corpus of sentential paraphrases},\n author={Dolan, William B and Brockett, Chris},\n booktitle={Proceedings of the Third International Workshop on Paraphrasing (IWP2005)},\n year={2005}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"text_a": {"dtype": "string", "id": null, "_type": "Value"}, "text_b": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "mrpc_da", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 845906, "num_examples": 3261, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 104536, "num_examples": 407, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 106918, "num_examples": 408, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/da/mrpc.tar.gz": {"num_bytes": 374973, "checksum": "d7b0520414351345f24660e26f2ae96011251908ac68533f876b16e14f904868"}}, "download_size": 374973, "post_processing_size": null, "dataset_size": 1057360, "size_in_bytes": 1432333}, "wnli_da": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Winograd Schema Challenge (Levesque et al., 2011) is a reading comprehension task\nin which a system must read a sentence with a pronoun and select the referent of that pronoun from\na list of choices. The examples are manually constructed to foil simple statistical methods: Each\none is contingent on contextual information provided by a single word or phrase in the sentence.\nTo convert the problem into sentence pair classification, we construct sentence pairs by replacing\nthe ambiguous pronoun with each possible referent. The task is to predict if the sentence with the\npronoun substituted is entailed by the original sentence. We use a small evaluation set consisting of\nnew examples derived from fiction books that was shared privately by the authors of the original\ncorpus. While the included training set is balanced between two classes, the test set is imbalanced\nbetween them (65% not entailment). Also, due to a data quirk, the development set is adversarial:\nhypotheses are sometimes shared between training and development examples, so if a model memorizes the\ntraining examples, they will predict the wrong label on corresponding development set\nexample. As with QNLI, each example is evaluated separately, so there is not a systematic correspondence\nbetween a model's score on this task and its score on the unconverted original task. We\ncall converted dataset WNLI (Winograd NLI).", "citation": "@inproceedings{levesque2012winograd,\n title={The winograd schema challenge},\n author={Levesque, Hector and Davis, Ernest and Morgenstern, Leora},\n booktitle={Thirteenth International Conference on the Principles of Knowledge Representation and Reasoning},\n year={2012}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "wnli_da", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 92825, "num_examples": 565, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 11391, "num_examples": 70, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 11753, "num_examples": 71, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/da/wnli.tar.gz": {"num_bytes": 29413, "checksum": "a97d74d9d2da304a18da4366534ae6c1de3f5fd4cc9a388504813568374af1ea"}}, "download_size": 29413, "post_processing_size": null, "dataset_size": 115969, "size_in_bytes": 145382}, "sst_da": {"description": "GLUE, the General Language Understanding Evaluation benchmark\n(https://gluebenchmark.com/) is a collection of resources for training,\nevaluating, and analyzing natural language understanding systems.\n\nThe Stanford Sentiment Treebank consists of sentences from movie reviews and\nhuman annotations of their sentiment. The task is to predict the sentiment of a\ngiven sentence. We use the two-way (positive/negative) class split, and use only\nsentence-level labels.", "citation": "@inproceedings{socher2013recursive,\n title={Recursive deep models for semantic compositionality over a sentiment treebank},\n author={Socher, Richard and Perelygin, Alex and Wu, Jean and Chuang, Jason and Manning, Christopher D and Ng, Andrew and Potts, Christopher},\n booktitle={Proceedings of the 2013 conference on empirical methods in natural language processing},\n pages={1631--1642},\n year={2013}\n}\n@article{wang2019superglue,\n title={SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems},\n author={Wang, Alex and Pruksachatkun, Yada and Nangia, Nikita and Singh, Amanpreet and Michael, Julian and Hill, Felix and Levy, Omer and Bowman, Samuel R},\n journal={arXiv preprint arXiv:1905.00537},\n year={2019}\n}\n\nNote that each SuperGLUE dataset has its own citation. Please see the source to\nget the correct citation for each contained dataset.\n", "homepage": "", "license": "", "features": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "string", "id": null, "_type": "Value"}, "idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "over_lim", "config_name": "sst_da", "version": {"version_str": "1.0.2", "description": null, "major": 1, "minor": 0, "patch": 2}, "splits": {"train": {"name": "train", "num_bytes": 4460298, "num_examples": 66486, "dataset_name": "over_lim"}, "validation": {"name": "validation", "num_bytes": 57489, "num_examples": 863, "dataset_name": "over_lim"}, "test": {"name": "test", "num_bytes": 104627, "num_examples": 872, "dataset_name": "over_lim"}}, "download_checksums": {"https://huggingface.co./datasets/KBLab/overlim/resolve/main/data/da/sst.tar.gz": {"num_bytes": 1929166, "checksum": "5fd539bff626886cf16d31223fa055b5abc3806684606e6e65a7978024fe96de"}}, "download_size": 1929166, "post_processing_size": null, "dataset_size": 4622414, "size_in_bytes": 6551580}} |