--- language: - en license: apache-2.0 tags: - text - question-and-answer pretty_name: AQuA-Rat task_categories: - question-answering --- # Dataset Card for AQuA-Rat - **Homepage:** [https://github.com/google-deepmind/AQuA](https://github.com/google-deepmind/AQuA) - This is an unofficial curation of the AQuA-Rat dataset, uploaded here with minimal (i.e., no content-modifying) processing. - **Paper:** [*Program Induction by Rationale Generation: Learning to Solve and Explain Algebraic Word Problems*](https://aclanthology.org/P17-1015.pdf) (ACL Anthology) Modifications: - Pre-tokenized splits removed since tokenization built into most LM pipelines. - Fixed file suffix from `.json` to `.jsonl`. - Changed `options` column to hardcoded A-E columns for easier parsing. - Changed `correct` column name to `answer_idx` and added `answer` containing the text of the correct answer. ## Dataset Structure Column names have been slightly modified from the original dataset: ``` question: The question/prompt. rationale: The explanation for the solution. A: The "A" answer. B: The "B" answer. C: The "C" answer. D: The "D" answer. E: The "E" answer. answer_idx: The multiple-choice identifier for the correct answer. answer: The correct answer. ``` Compare this example with its equivalent from the original dataset. New: ```json { "question": "Find out which of the following values is the multiple of X, if it is divisible by 9 and 12?", "rationale": "9=3*3\n12=3*4\nThe number should definitely have these factors 3*3*4\n36 is the number that has these factors\nSo, 36 is the multiple of X\nAnswer is A", "A": "36", "B": "15", "C": "17", "D": "5", "E": "7", "answer_idx": "A", "answer": "36" } ``` Original: ```json { "question": "Find out which of the following values is the multiple of X, if it is divisible by 9 and 12?", "options": ["A)36", "B)15", "C)17", "D)5", "E)7"], "rationale": "9=3*3\n12=3*4\nThe number should definitely have these factors 3*3*4\n36 is the number that has these factors\nSo, 36 is the multiple of X\nAnswer is A", "correct": "A", } ``` No other changes were made. ## Citation Information For reproducibility, please include a link to *this* dataset when publishing results based on the included data. For formal citations, please cite the *original* publication: ``` @inproceedings{ling-etal-2017-program, title = "Program Induction by Rationale Generation: Learning to Solve and Explain Algebraic Word Problems", author = "Ling, Wang and Yogatama, Dani and Dyer, Chris and Blunsom, Phil", editor = "Barzilay, Regina and Kan, Min-Yen", booktitle = "Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)", month = jul, year = "2017", address = "Vancouver, Canada", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/P17-1015/", doi = "10.18653/v1/P17-1015", pages = "158--167" } ```