five

iknow-lab/wildguardmix-train-ko-11k

收藏
Hugging Face2024-10-19 更新2025-04-12 收录
下载链接:
https://hf-mirror.com/datasets/iknow-lab/wildguardmix-train-ko-11k
下载链接
链接失效反馈
官方服务:
资源简介:
--- dataset_info: features: - name: prompt dtype: string - name: adversarial dtype: bool - name: response dtype: string - name: prompt_harm_label dtype: string - name: response_refusal_label dtype: string - name: response_harm_label dtype: string - name: subcategory dtype: string - name: prompt_ko dtype: string - name: response_ko dtype: string splits: - name: train num_bytes: 48289970 num_examples: 11105 download_size: 24569189 dataset_size: 48289970 configs: - config_name: default data_files: - split: train path: data/train-* license: odc-by task_categories: - text-classification language: - ko size_categories: - 10K<n<100K --- - Original dataset: [allenai/wildguardmix](https://huggingface.co/datasets/allenai/wildguardmix) - Translated by [nayohan/llama3-instrucTrans-enko-8b](https://huggingface.co/nayohan/llama3-instrucTrans-enko-8b) ``` @misc{wildguard2024, title={WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs}, author={Seungju Han and Kavel Rao and Allyson Ettinger and Liwei Jiang and Bill Yuchen Lin and Nathan Lambert and Yejin Choi and Nouha Dziri}, year={2024}, eprint={2406.18495}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2406.18495}, } ```

数据集信息: 特征项: - 名称:提示词(prompt),数据类型:字符串 - 名称:对抗性标识(adversarial),数据类型:布尔值 - 名称:回复内容(response),数据类型:字符串 - 名称:提示词危害标签(prompt_harm_label),数据类型:字符串 - 名称:回复拒答标签(response_refusal_label),数据类型:字符串 - 名称:回复危害标签(response_harm_label),数据类型:字符串 - 名称:子类别(subcategory),数据类型:字符串 - 名称:韩语版提示词(prompt_ko),数据类型:字符串 - 名称:韩语版回复内容(response_ko),数据类型:字符串 数据划分: - 划分名称:训练集(train),字节占用量:48289970,样本总数:11105 整体下载大小:24569189,数据集总存储大小:48289970 配置项: - 配置名称:默认配置(default),数据文件信息: - 对应训练集划分的数据路径:data/train-* 开源协议:odc-by 任务类别:文本分类(text-classification) 数据集语言:韩语(ko) 样本规模区间:10K < n < 100K - 原始数据集:[allenai/wildguardmix](https://huggingface.co/datasets/allenai/wildguardmix) - 翻译工具:[nayohan/llama3-instrucTrans-enko-8b](https://huggingface.co/nayohan/llama3-instrucTrans-enko-8b) @misc{wildguard2024, title={WildGuard:面向大语言模型(LLM)安全风险、越狱攻击与模型拒答行为的一站式开源审核工具}, author={Seungju Han and Kavel Rao and Allyson Ettinger and Liwei Jiang and Bill Yuchen Lin and Nathan Lambert and Yejin Choi and Nouha Dziri}, year={2024}, eprint={2406.18495}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2406.18495}, }
提供机构:
iknow-lab
二维码
社区交流群
二维码
科研交流群
商业服务