Elynden/AgentBench-EvoSyn

本数据集包含使用EvoSyn框架合成的高质量操作系统智能体任务，每个任务包括任务描述、系统初始化脚本和区分性测试脚本。这些操作系统任务分为两种类型：一种需要模型提供最终结果，即QA类型；另一种需要模型完成任务，即EXEC类型。不同类型的测试逻辑各不相同。

This dataset contains high-quality operating system agent tasks synthesized and filtered using the EvoSyn framework, including task description, system initialization scripts, and discriminative test scripts. These OS tasks are divided into two types: one requires the model to provide a final result, which is the QA type, and the other requires the model to complete a task, which is the EXEC type. The logic of different types of testing varies.