LLM-Artifacts

该数据集是由大型语言模型（LLM）生成的人工数据集，旨在为多种应用场景提供支持，包括分类任务、偏好评估、指令提示、模拟对话以及自由文本等形式。该数据集包含五种根据输出限制分类的数据类型，其目的是为了更深入地理解LLM的优势和局限性。这些数据集涉及的任务包括但不限于分类、人类对齐、指令微调、社交互动模拟以及数据稀缺补偿。

This is an artificial dataset generated by Large Language Models (LLMs), designed to support diverse application scenarios including classification tasks, preference evaluation, instruction prompting, simulated dialogue, and free-text formats. This dataset encompasses five types of data categorized based on output constraints, with the goal of enabling a deeper understanding of the strengths and limitations of LLMs. The tasks covered by this dataset include, but are not limited to, classification, human alignment, instruction fine-tuning, social interaction simulation, and data scarcity compensation.