agentlans/prompt-difficulty

该数据集专注于评估大型语言模型（LLMs）提示的难度。它详细介绍了评估提示难度的方法，包括从特定数据集中选择提示、用于评估的模型以及难度评估的标准。结果部分展示了不同模型在难度评分上的一致性，并提供了提示示例及其对应的难度分数。此外，还讨论了该方法的局限性和结论，强调了难度度量在课程学习、数据集过滤和性能预测等方面的潜在应用。

This dataset focuses on assessing the difficulty of prompts for large language models (LLMs). It details the method used to evaluate prompt difficulty, including the selection of prompts from a specific dataset, the models used for evaluation, and the criteria for difficulty assessment. The results section highlights the consistency of difficulty ratings across models and provides examples of prompts with their corresponding difficulty scores. Limitations and conclusions are also discussed, emphasizing the potential applications of the difficulty metric in areas such as curriculum learning, dataset filtering, and performance prediction.