mrm8488/unnatural-instructions-core
收藏资源简介：
--- dataset_info: features: - name: instruction dtype: string - name: instances list: - name: instruction_with_input dtype: string - name: input dtype: string - name: constraints dtype: string - name: output dtype: string splits: - name: train num_bytes: 54668900 num_examples: 66010 download_size: 28584196 dataset_size: 54668900 --- # Dataset Card for Unnatural Instructions (Core data) This info comes from the **Unnatural Instructions GitHub [repo](https://github.com/orhonovich/unnatural-instructions/)**. Unnatural Instructions is a dataset of instructions automatically generated by a Large Language model. See full details in the paper: "[Unnatural Instructions: Tuning Language Models with (Almost) No Human Labor](https://arxiv.org/abs/2212.09689)" ## 🗃️ Content The Unnatural Instructions core dataset of 68,478 instruction-input-output triplets. ## 📄 Format ### Core data Each example contains: - `input`: An input for the task described by the `instruction` - `instruction_with_input`: The instruction concatenated with the `input` - `constraints`: The task's output space constraints - `output`: The output of executing `instruction` with the given `input` ## 📘 Citation If you make use of Unnatural Instructions, please cite the following paper: ``` @misc{honovich2022unnatural, title = {Unnatural Instructions: Tuning Language Models with (Almost) No Human Labor}, author = {Honovich, Or and Scialom, Thomas and Levy, Omer and Schick, Timo}, url = {https://arxiv.org/abs/2212.09689}, publisher = {arXiv}, year={2022} } ``` [More Information needed](https://github.com/huggingface/datasets/blob/main/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
数据集信息： 特征： - 名称：instruction，数据类型：字符串 - 名称：instances，类型为列表，包含以下子字段： - 名称：instruction_with_input，数据类型：字符串 - 名称：input，数据类型：字符串 - 名称：constraints，数据类型：字符串 - 名称：output，数据类型：字符串 划分集： - 名称：train（训练集），字节数：54668900，样本数：66010 下载大小：28584196，数据集总大小：54668900 # 非自然指令（核心数据）数据集卡片 本信息源自**非自然指令GitHub仓库[repo](https://github.com/orhonovich/unnatural-instructions/)**。 非自然指令（Unnatural Instructions）是一个由大语言模型（Large Language Model）自动生成的指令数据集。详细信息请参阅论文：《非自然指令：使用（几乎）零人类人力微调语言模型》（https://arxiv.org/abs/2212.09689） ## 🗃️ 数据集内容 非自然指令核心数据集包含68478条指令-输入-输出三元组样本。 ## 📄 数据格式 ### 核心数据格式 每条样本包含以下字段： - `input`：对应`instruction`所描述任务的输入内容 - `instruction_with_input`：将`instruction`与`input`拼接而成的完整指令 - `constraints`：该任务的输出空间约束规则 - `output`：针对给定`input`执行`instruction`所得到的输出结果 ## 📘 引用规范 若您使用非自然指令数据集，请引用以下论文： @misc{honovich2022unnatural, title = {Unnatural Instructions: Tuning Language Models with (Almost) No Human Labor}, author = {Honovich, Or and Scialom, Thomas and Levy, Omer and Schick, Timo}, url = {https://arxiv.org/abs/2212.09689}, publisher = {arXiv}, year={2022} } [更多信息请参考](https://github.com/huggingface/datasets/blob/main/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
数据集卡片：Unnatural Instructions（核心数据）
数据集信息
-
特征:
instruction: 字符串类型instances: 列表类型，包含以下子项:instruction_with_input: 字符串类型input: 字符串类型constraints: 字符串类型output: 字符串类型
-
拆分:
train:- 字节数: 54668900
- 样本数: 66010
-
下载大小: 28584196 字节
-
数据集大小: 54668900 字节
内容
Unnatural Instructions 核心数据集包含 68,478 个指令-输入-输出三元组。
格式
核心数据
每个示例包含:
input: 任务的输入instruction_with_input: 指令与输入的连接constraints: 任务输出空间的约束output: 执行指令并使用给定输入的输出
引用
如果您使用 Unnatural Instructions 数据集，请引用以下论文:
@misc{honovich2022unnatural, title = {Unnatural Instructions: Tuning Language Models with (Almost) No Human Labor}, author = {Honovich, Or and Scialom, Thomas and Levy, Omer and Schick, Timo}, url = {https://arxiv.org/abs/2212.09689}, publisher = {arXiv}, year={2022} }
