CAMA: On Benchmarking Code LLMs for Android Malware Analysis

CAMA This dataset is published as part of the paper "CAMA: On Benchmarking Code LLMs for Android Malware Analysis", which has been accepted to the 34th ACM SIGSOFT ISSTA Companion (LLMSC Workshop 2025). It includes the structured model outputs specified by Cama, consisting of function summaries, refined function names, and maliciousness scores. The dataset folder contains the output results of Cama on 118 Android malware samples for four popular open-source models, covering over 7.5 million distinct functions. For more details, please refer to the original paper: https://arxiv.org/abs/2504.00694.CAMA 本数据集作为论文《CAMA：面向安卓恶意软件分析的代码大语言模型（Code LLMs）基准测试》（CAMA: On Benchmarking Code LLMs for Android Malware Analysis）的附属研究成果发布，该论文已被第34届ACM SIGSOFT国际软件测试与分析大会附属研讨会（LLMSC Workshop 2025）收录。数据集包含CAMA定义的结构化模型输出结果，涵盖函数摘要、精炼后的函数名称以及恶意性评分。 数据集文件夹中收录了CAMA针对118个安卓恶意软件样本，在四款主流开源代码大语言模型上的运行输出结果，覆盖超过750万个独立函数。 如需了解更多细节，请参阅原论文：https://arxiv.org/abs/2504.00694。