sKT-Ai-Labs/HQ-QUESTION-PAIRS
收藏资源简介:
该数据集包含由高性能大型语言模型生成的大量问题对。它专门为希望评估模型在问候复杂性、深度思考、逻辑推理和复杂编码任务中能力的研究人员和开发人员设计。这是一个精心策划的高复杂度问题对集合,旨在将大型语言模型推向极限。不包含答案,旨在用于直接生成、评估和系统提示调优。数据集涵盖多领域,从简单的问候复杂性到高级算法编码挑战,并针对系统提示优化,包含数千对问题,适用于大规模基准测试。
This dataset contains a large number of question pairs generated by high-performance large language models. It is specifically designed for researchers and developers who wish to evaluate model capabilities across scenarios including greeting complexity, deep thinking, logical reasoning, and complex coding tasks. As a curated collection of high-complexity question pairs, it is intended to push large language models to their limits. The dataset includes no answers, and is tailored for direct generation, model evaluation, and system prompt tuning. Covering multiple domains ranging from simple greeting-related complexity scenarios to advanced algorithmic coding challenges, this dataset has been optimized for system prompts, contains thousands of question pairs, and is suitable for large-scale benchmarking.




