D130_riddles_and_brain_teasers_dataset

本数据集精心收集超过10万条中文谜语及3000余条脑筋急转弯，涵盖字谜、物谜、智力急转弯等多种趣味题型。数据经过清洗与结构化，专用于训练大模型理解语言双关、隐喻、发散思维与创意推理能力，是开发智能对话、娱乐互动AI及评测模型幽默感的优质资源。

This dataset has carefully curated over 100,000 Chinese riddles and more than 3,000 brain teasers, covering various engaging question types including word puzzles, object riddles, and intellectual brain teasers. The data has been cleaned and structured, and is specifically designed to train large language models (LLMs) in understanding linguistic puns, metaphors, divergent thinking and creative reasoning abilities. It serves as a premium resource for developing intelligent dialogue systems, entertainment-oriented interactive AI, and evaluating models' sense of humor.