Benchmark Dataset for Cognitive Biases in LLMs

该数据集是一个包含30,000项测试的基准数据集，旨在检测大型语言模型中的认知偏见。它是针对大型语言模型的一种新颖通用测试框架的一部分，该框架能够可靠地生成测试。该数据集的规模为30,000项测试，其任务是评估大型语言模型中的认知偏见。

This is a benchmark dataset containing 30,000 test instances, whose core task is to detect and evaluate cognitive biases in large language models (LLMs). It is part of a novel general-purpose test framework tailored for LLMs, which can reliably generate standardized test cases.