Situation Puzzle

该数据集是一个包含975个分级情境谜题的基准测试，这些谜题分为三个难度级别，旨在评估和引导大型语言模型（LLM）的横向思维。这些谜题被分为简单、中等和困难三个难度类别，并通过网络爬虫过程收集，之后由人工审核以确保质量。该数据集规模包含975个独特的情境谜题，其任务是评估大型语言模型的横向思维能力。

This dataset is a benchmark consisting of 975 graded situational puzzles, which are divided into three difficulty levels and designed to evaluate and elicit lateral thinking abilities in large language models (LLMs). These puzzles are categorized into three difficulty classes: easy, medium, and hard. They were collected via web crawling and subsequently manually reviewed to ensure quality. This dataset contains 975 unique situational puzzles, whose task is to evaluate the lateral thinking capabilities of large language models.