cmu-lti/cobracorpus
收藏资源简介:
--- license: openrail task_categories: - text-generation - text-classification language: - en pretty_name: COBRA🐍 size_categories: - 10K<n<100K --- # Dataset Card for Dataset Name ## Dataset Description - **Homepage: https://cobra.xuhuiz.com/** - **Paper: https://arxiv.org/abs/2306.01985** ### Dataset Summary This dataset contains COBRACOPURS and COBRACORPUS-counterfactual in this [paper](https://arxiv.org/abs/2306.01985) ### Data Splits * `advContexts_explanations.csv` is `COBRACorpus-CF` * `toxigen_explanations.csv` is the full `COBRACorpus` * `toxigen_explanations_train.csv` is the training split of `COBRACorpus` * `toxigen_explanations_val.csv` is the validation split of `COBRACorpus` ### Data Entries For `COBRACorpus`, the relevant entries in the `csv` files are *`situationalContext (string)`, `speakerIdentity (string)`, `listenerIdentity (string)`, `statement (string)`, `intent (string)`, `targetGroup (string)`, `relevantPowerDynamics (string)`, `implication (string)`, `targetGroupEmotionalReaction (string)`, `targetGroupCognitiveReaction (string)`, `offensiveness (string)`* Please refer to the [paper](https://arxiv.org/abs/2306.01985) for the specific explanations of these entries. The *`examples`* entry is the few-shot prompt that we used to generate explanations. All other entries are from the [Toxicgen](https://arxiv.org/abs/2203.09509) dataset, which is not directly relevant to this work but we leave them there as the metadata in case it's useful for the future works. ### Citation Information If you find this dataset useful, please cite: ``` @inproceedings{zhou2023cobra, title = {COBRA Frames: Contextual Reasoning about Effects and Harms of Offensive Statements}, author = {Zhou, Xuhui and Zhu, Hao and Yerukola, Akhila and Davidson, Thomas and D. Hwang, Jena and Swayamdipta, Swabha and Sap, Maarten}, year = {2023}, booktitle = {Findings of ACL} } ```
license: openrail 任务类别: - 文本生成(text-generation) - 文本分类(text-classification) 语言: - 英语(en) 展示名称:COBRA🐍 规模类别: - 10,000 < 样本数量 < 100,000 --- # 数据集卡片 ## 数据集说明 - **主页:https://cobra.xuhuiz.com/** - **相关论文:https://arxiv.org/abs/2306.01985** ### 数据集概述 本数据集涵盖本论文中提及的COBRACOPURS与反事实COBRACORPUS(COBRACORPUS-counterfactual)。 ### 数据划分 * `advContexts_explanations.csv` 对应 `COBRACorpus-CF` * `toxigen_explanations.csv` 为完整的 `COBRACorpus` * `toxigen_explanations_train.csv` 为 `COBRACorpus` 的训练集划分 * `toxigen_explanations_val.csv` 为 `COBRACorpus` 的验证集划分 ### 数据条目 针对 `COBRACorpus`,其CSV文件中的有效条目包括: * 情境上下文(situationalContext,字符串类型)、说话者身份(speakerIdentity,字符串类型)、倾听者身份(listenerIdentity,字符串类型)、陈述内容(statement,字符串类型)、意图(intent,字符串类型)、目标群体(targetGroup,字符串类型)、相关权力动态(relevantPowerDynamics,字符串类型)、隐含含义(implication,字符串类型)、目标群体情绪反应(targetGroupEmotionalReaction,字符串类型)、目标群体认知反应(targetGroupCognitiveReaction,字符串类型)、冒犯性程度(offensiveness,字符串类型) 有关这些条目的详细说明,请参阅本论文。 `examples` 条目为我们用于生成解释的少样本提示(few-shot prompt)。 其余所有条目均来自[Toxicgen](https://arxiv.org/abs/2203.09509)数据集,该数据集与本研究无直接关联,但我们保留其作为元数据,以供后续研究参考使用。 ### 引用信息 若您认为本数据集对研究有所帮助,请引用如下文献: @inproceedings{zhou2023cobra, title = {COBRA框架:冒犯性陈述的影响与危害的上下文推理}, author = {Zhou, Xuhui and Zhu, Hao and Yerukola, Akhila and Davidson, Thomas and D. Hwang, Jena and Swayamdipta, Swabha and Sap, Maarten}, year = {2023}, booktitle = {ACL 研究发现} }
数据集概述
基本信息
- 许可证: openrail
- 任务类别:
- 文本生成
- 文本分类
- 语言: 英语
- 数据集名称: COBRA🐍
- 数据集大小: 10K<n<100K
数据集详情
- 数据集总结: 包含COBRACOPURS和COBRACORPUS-counterfactual,详细信息参考论文。
- 数据分割:
advContexts_explanations.csv: COBRACorpus-CFtoxigen_explanations.csv: 完整的COBRACorpustoxigen_explanations_train.csv: COBRACorpus的训练集toxigen_explanations_val.csv: COBRACorpus的验证集
- 数据条目:
- 对于
COBRACorpus,相关条目包括:situationalContext,speakerIdentity,listenerIdentity,statement,intent,targetGroup,relevantPowerDynamics,implication,targetGroupEmotionalReaction,targetGroupCognitiveReaction,offensiveness。 examples条目是用于生成解释的少样本提示。
- 对于
引用信息
若使用此数据集,请引用:
@inproceedings{zhou2023cobra, title = {COBRA Frames: Contextual Reasoning about Effects and Harms of Offensive Statements}, author = {Zhou, Xuhui and Zhu, Hao and Yerukola, Akhila and Davidson, Thomas and D. Hwang, Jena and Swayamdipta, Swabha and Sap, Maarten}, year = {2023}, booktitle = {Findings of ACL} }




