营销知识评测数据集V1.0

营销知识评测数据集是专为AI营销模型训练与评估设计的开放式标准数据集，现已正式开源。该数据集共包含600余道经领域专家审核的多维度中文评测题，覆盖品牌认知、产品定位、营销策略等核心知识点。题目统一采用“问题 + 选项 + 标准答案选项”的结构形式，通过限定答案范围以提升评测的一致性与可量化性，便于模型能力的精准对比与持续优化。本数据集适用于大模型预训练后下游任务的适配评估，也可作为营销类AI应用的基准测试工具。

The Marketing Knowledge Evaluation Dataset is an open standard dataset specifically designed for training and evaluating AI marketing models, and it has been officially open-sourced. This dataset contains over 600 multi-dimensional Chinese evaluation questions audited by domain experts, covering core knowledge points such as brand awareness, product positioning, and marketing strategies. All questions uniformly adopt the format of "Question + Options + Correct Answer Option". By limiting the answer scope, the consistency and quantifiability of the evaluation are enhanced, which facilitates accurate comparison and continuous optimization of model capabilities. This dataset is suitable for adaptive evaluation of downstream tasks after pre-training of large language models, and can also serve as a benchmark testing tool for marketing-oriented AI applications.

搜集汇总

背景与挑战

背景概述

该数据集由深圳东信云开发，包含1300多条经过专家评审的营销知识评测数据，专为AI营销模型的训练和评估设计，涵盖品牌认知、产品定位和营销策略等多个核心维度。所有问题采用统一的'问题+选项+标准答案'格式，便于量化比较和模型优化，适用于评估大语言模型在预训练后对下游任务的适应能力，也可作为AI在营销领域应用的基准测试工具。