tasksource/winodict

--- language: en license: cc-by-4.0 dataset_info: features: - name: id dtype: string - name: lemma dtype: string - name: fake_lemma dtype: string - name: pos dtype: string - name: tag dtype: string - name: pronoun dtype: string - name: definition dtype: string - name: sentence dtype: string - name: option1 dtype: string - name: option2 dtype: string - name: label dtype: int64 splits: - name: train num_bytes: 415190 num_examples: 1488 - name: val num_bytes: 135624 num_examples: 496 - name: test num_bytes: 135191 num_examples: 496 download_size: 249676 dataset_size: 686005 --- https://github.com/google-research/language/tree/master/language/wino_dict ```@inproceedings{51779, title = {WinoDict: Probing language models for in-context language acquisition}, author = {Fangyu Liu and Jeremy Cole and Julian Martin Eisenschlos and William Weston Cohen}, year = {2022}, URL = {https://arxiv.org/abs/2209.12153}, booktitle = {EACL} } ```

原始信息汇总

数据集概述

数据集特征

• id: 字符串类型
• lemma: 字符串类型
• fake_lemma: 字符串类型
• pos: 字符串类型
• tag: 字符串类型
• pronoun: 字符串类型
• definition: 字符串类型
• sentence: 字符串类型
• option1: 字符串类型
• option2: 字符串类型
• label: 整数类型（int64）

数据集划分

• train: 1488个样本，大小为415190字节
• val: 496个样本，大小为135624字节
• test: 496个样本，大小为135191字节

数据集大小

• 下载大小: 249676字节
• 数据集总大小: 686005字节

许可证

• 许可证类型: CC-BY-4.0