Convai

Convai数据集是由赫瑞瓦特大学的交互实验室创建，专注于研究对话AI系统中的滥用语言。该数据集包含超过6000个样本，覆盖了三种不同类型的对话AI系统，旨在通过多专家注释反映滥用语言的细微差别。数据集主要用于开发和评估滥用检测模型，特别是在性别歧视和性骚扰方面的检测，以帮助改进对话AI系统的安全性和用户体验。

The Convai Dataset was created by the Interaction Lab at Heriot-Watt University, focusing on the study of abusive language in conversational AI systems. This dataset contains over 6,000 samples covering three distinct types of conversational AI systems, and aims to reflect the nuances of abusive language via multi-expert annotation. It is primarily used for developing and evaluating abusive language detection models, particularly those targeting gender discrimination and sexual harassment, to help improve the safety and user experience of conversational AI systems.

搜集汇总

背景与挑战

背景概述

Convai数据集是由赫瑞瓦特大学交互实验室创建的对话AI滥用语言研究资源，包含超过6000个样本，覆盖三种对话AI系统，通过多专家注释捕捉滥用语言的细微差别。该数据集主要用于开发和评估滥用检测模型，特别是在性别歧视和性骚扰方面，旨在提升对话AI系统的安全性和用户体验。