RANEPA-ai/SLAVA-OpenData-2800-v1

SLAVA是一个用于评估大型语言模型（LLM）在俄罗斯领域事实准确性的基准。该数据集包含多种类型的问题，用于测试LLM的知识准确性、问题的敏感性和创建基于答案准确性的综合评估系统。数据集的问题分为单选、多选和开放回答三种类型，并针对问题的敏感性进行评分。测试结果显示，GigaChat、YandexGPT和qwen2模型在准确性和处理复杂、挑衅性问题方面表现出色。

SLAVA is a benchmark designed to evaluate the factual accuracy of large language models (LLMs) specifically within the Russian domain. The dataset contains various types of questions to test the knowledge accuracy of LLMs, the sensitivity of the questions, and to create a comprehensive evaluation system based on answer accuracy. The questions in the dataset are divided into three types: single choice, multiple choice, and open-ended responses, and are rated for their sensitivity. The test results show that models from GigaChat, YandexGPT, and qwen2 have demonstrated high accuracy and the ability to handle complex, provocative questions.