cardiffnlp/relentless

--- language: - en license: - other multilinguality: - monolingual size_categories: - n<1K pretty_name: relentless --- # Dataset Card for "cardiffnlp/relentless" ***RelEntLess*** is a new benchmark, in which entity pairs have to be ranked according to how much they satisfy a given graded relation. Essentially, the task is a ranking task where we provide five prototypical examples to each relation. Following brief description of each relation type is used in our baseline in addition to the prototypical examples. Please check our paper "[A RelEntLess Benchmark for Modelling Graded Relations between Named Entities](https://arxiv.org/abs/2305.15002)" for more detail. ```python { "friend/ally of": "entities that are friends or allies", "competitor/rival of": "entities that are competitors or rivals", "known for": "examples of what entities are known for", "influenced by": "what has influenced different entities", "similar to": "examples of entities that are similar" } ``` ## Dataset Description - **Repository:** [https://huggingface.co/datasets/cardiffnlp/relentless](https://huggingface.co/datasets/cardiffnlp/relentless) - **Paper:** [A RelEntLess Benchmark for Modelling Graded Relations between Named Entities](https://arxiv.org/abs/2305.15002) - **Dataset:** [https://huggingface.co/datasets/cardiffnlp/relentless](https://huggingface.co/datasets/cardiffnlp/relentless) ### Dataset Summary | relation_type | val. | test | |:--------------------|-------:|-------:| | competitor/rival of | 20 | 84 | | friend/ally of | 19 | 88 | | influenced by | 19 | 90 | | known for | 18 | 105 | | similar to | 19 | 89 | ## Dataset Structure ### Data Instances ```python { "pairs": [["Le Corbusier", "purism art"], ["Sean Connery", "Finding Forrester"], ...], "scores_all": [[4.0, 5.0, 3.0, 4.0, 5.0, 3.0, 5.0], [4.0, 5.0, 2, 5.0, 5.0, 4.0, 2], ...], "scores_mean": [4.142857142857143, 3.857142857142857, 4.857142857142857, ...], "relation_type": "known for", "ranks": [8.5, 11, 5, 14, 15, 5, 20, 13, 1.5, 18, 10, 1.5, 17, ...], "prototypical_examples": [ [ "Russell Crowe", "Gladiator" ], [ "Cadbury", "chocolate" ],...] } ``` ### Citation Information ``` @misc{ushio2023relentless, title={A RelEntLess Benchmark for Modelling Graded Relations between Named Entities}, author={Asahi Ushio and Jose Camacho Collados and Steven Schockaert}, year={2023}, eprint={2305.15002}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```

原始信息汇总

数据集卡片 "cardiffnlp/relentless"

数据集描述

• 语言: 英语
• 许可证: 其他
• 多语言性: 单语
• 大小类别: n<1K
• 名称: relentless

数据集概述

RelEntLess 是一个新的基准数据集，其中实体对需要根据它们满足给定分级关系的程度进行排序。本质上，这是一个排序任务，我们为每个关系类型提供五个原型示例。此外，每个关系类型的简要描述用于我们的基线模型中，除了原型示例外。

数据集结构

数据实例

python { "pairs": [["Le Corbusier", "purism art"], ["Sean Connery", "Finding Forrester"], ...], "scores_all": [[4.0, 5.0, 3.0, 4.0, 5.0, 3.0, 5.0], [4.0, 5.0, 2, 5.0, 5.0, 4.0, 2], ...], "scores_mean": [4.142857142857143, 3.857142857142857, 4.857142857142857, ...], "relation_type": "known for", "ranks": [8.5, 11, 5, 14, 15, 5, 20, 13, 1.5, 18, 10, 1.5, 17, ...], "prototypical_examples": [ [ "Russell Crowe", "Gladiator" ], [ "Cadbury", "chocolate" ],...] }

关系类型统计

relation_type	val.	test
competitor/rival of	20	84
friend/ally of	19	88
influenced by	19	90
known for	18	105
similar to	19	89

引用信息

@misc{ushio2023relentless, title={A RelEntLess Benchmark for Modelling Graded Relations between Named Entities}, author={Asahi Ushio and Jose Camacho Collados and Steven Schockaert}, year={2023}, eprint={2305.15002}, archivePrefix={arXiv}, primaryClass={cs.CL} }