COMPL-AI Benchmarking Suite
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/compl-ai/compl-ai
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一个旨在评估大型语言模型（LLM）在欧盟人工智能法案背景下的基准测试套件，它重点关注伦理原则和技术要求。该套件包含23个基准测试，涵盖了透明度、多样性、非歧视和公平性等不同的伦理原则。本次评估涵盖了12个知名LLM，任务是对这些模型与欧盟人工智能法案要求的一致性进行基准测试。
This dataset is a benchmark suite designed to evaluate Large Language Models (LLMs) against the requirements of the European Union Artificial Intelligence Act, with a focus on ethical principles and technical specifications. The suite includes 23 benchmark tasks covering diverse ethical principles such as transparency, diversity, non-discrimination, and fairness. This evaluation covers 12 prominent LLMs, aiming to benchmark the compliance of these models with the provisions of the EU Artificial Intelligence Act.
提供机构：
COMPL-AI framework authors
