arc_challenge-hellaswag-piqa

arc_challenge-hellaswag-piqa
-----
{"description": "", "citation": "", "homepage": "", "license": "", "features": {"id": {"dtype": "string", "_type": "Value"}, "question": {"dtype": "string", "_type": "Value"}, "choices": {"feature": {"text": {"dtype": "string", "_type": "Value"}, "label": {"dtype": "string", "_type": "Value"}}, "_type": "Sequence"}, "answerKey": {"dtype": "string", "_type": "Value"}}, "builder_name": "parquet", "dataset_name": "ai2_arc", "config_name": "ARC-Challenge", "version": {"version_str": "0.0.0", "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 349760, "num_examples": 1119, "dataset_name": "ai2_arc"}, "test": {"name": "test", "num_bytes": 375511, "num_examples": 1172, "dataset_name": "ai2_arc"}, "validation": {"name": "validation", "num_bytes": 96660, "num_examples": 299, "dataset_name": "ai2_arc"}}, "download_checksums": {"hf://datasets/allenai/ai2_arc@210d026faf9955653af8916fad021475a3f00453/ARC-Challenge/train-00000-of-00001.parquet": {"num_bytes": 189909, "checksum": null}, "hf://datasets/allenai/ai2_arc@210d026faf9955653af8916fad021475a3f00453/ARC-Challenge/test-00000-of-00001.parquet": {"num_bytes": 203808, "checksum": null}, "hf://datasets/allenai/ai2_arc@210d026faf9955653af8916fad021475a3f00453/ARC-Challenge/validation-00000-of-00001.parquet": {"num_bytes": 55743, "checksum": null}}, "download_size": 449460, "dataset_size": 821931, "size_in_bytes": 1271391}
----
{"description": "\nHellaSwag: Can a Machine Really Finish Your Sentence? is a new dataset for commonsense NLI. A paper was published at ACL2019.\n", "citation": "@inproceedings{zellers2019hellaswag,\n title={HellaSwag: Can a Machine Really Finish Your Sentence?},\n author={Zellers, Rowan and Holtzman, Ari and Bisk, Yonatan and Farhadi, Ali and Choi, Yejin},\n booktitle ={Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics},\n year={2019}\n}\n", "homepage": "https://rowanzellers.com/hellaswag/", "license": "", "features": {"ind": {"dtype": "int32", "_type": "Value"}, "activity_label": {"dtype": "string", "_type": "Value"}, "ctx_a": {"dtype": "string", "_type": "Value"}, "ctx_b": {"dtype": "string", "_type": "Value"}, "ctx": {"dtype": "string", "_type": "Value"}, "endings": {"feature": {"dtype": "string", "_type": "Value"}, "_type": "Sequence"}, "source_id": {"dtype": "string", "_type": "Value"}, "split": {"dtype": "string", "_type": "Value"}, "split_type": {"dtype": "string", "_type": "Value"}, "label": {"dtype": "string", "_type": "Value"}}, "builder_name": "hellaswag", "dataset_name": "hellaswag", "config_name": "default", "version": {"version_str": "0.1.0", "major": 0, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 43232624, "num_examples": 39905, "dataset_name": "hellaswag"}, "test": {"name": "test", "num_bytes": 10791853, "num_examples": 10003, "dataset_name": "hellaswag"}, "validation": {"name": "validation", "num_bytes": 11175717, "num_examples": 10042, "dataset_name": "hellaswag"}}, "download_checksums": {"https://raw.githubusercontent.com/rowanz/hellaswag/master/data/hellaswag_train.jsonl": {"num_bytes": 47496131, "checksum": null}, "https://raw.githubusercontent.com/rowanz/hellaswag/master/data/hellaswag_test.jsonl": {"num_bytes": 11752147, "checksum": null}, "https://raw.githubusercontent.com/rowanz/hellaswag/master/data/hellaswag_val.jsonl": {"num_bytes": 12246618, "checksum": null}}, "download_size": 71494896, "dataset_size": 65200194, "size_in_bytes": 136695090}
----
{"description": "", "citation": "", "homepage": "", "license": "", "features": {"goal": {"dtype": "string", "_type": "Value"}, "sol1": {"dtype": "string", "_type": "Value"}, "sol2": {"dtype": "string", "_type": "Value"}, "label": {"names": ["0", "1"], "_type": "ClassLabel"}}, "builder_name": "parquet", "dataset_name": "piqa", "config_name": "plain_text", "version": "1.1.0", "splits": {"train": {"name": "train", "num_bytes": 4106017, "num_examples": 16113, "dataset_name": "piqa"}, "test": {"name": "test", "num_bytes": 761895, "num_examples": 3084, "dataset_name": "piqa"}, "validation": {"name": "validation", "num_bytes": 464539, "num_examples": 1838, "dataset_name": "piqa"}}, "download_checksums": {"hf://datasets/piqa@40bca61d846aee3fd71ee3617c6e835c053e803f/plain_text/train/0000.parquet": {"num_bytes": 2657677, "checksum": null}, "hf://datasets/piqa@40bca61d846aee3fd71ee3617c6e835c053e803f/plain_text/test/0000.parquet": {"num_bytes": 501652, "checksum": null}, "hf://datasets/piqa@40bca61d846aee3fd71ee3617c6e835c053e803f/plain_text/validation/0000.parquet": {"num_bytes": 301200, "checksum": null}}, "download_size": 3460529, "dataset_size": 5332451, "size_in_bytes": 8792980}

搜集汇总

背景与挑战

背景概述

该数据集整合了三个子数据集：ARC-Challenge（包含2590个问答样本）、HellaSwag（包含59950个常识推理样本）和PIQA（包含21035个物理常识问答样本）。这些子数据集分别用于评估模型在科学问题回答、句子补全和物理交互理解方面的能力。