Combining Large Language Models for High-quality, Cost-efficient Conservative Reassessments of Static Security Findings 2026

Anonymous review artifact accompanying the submission “Combining Large Language Models for High-quality, Cost-efficient Conservative Reassessments of Static Security Findings”. The archive includes the reused benchmark data, prompts, experiment outputs, evaluation notebooks, and figures required to inspect and reproduce the reported results. Author-identifying metadata and local environment traces were removed for double-blind review.本匿名评审附属数据集配套投稿论文《结合大语言模型（Large Language Model）实现静态安全发现结果的高质量、低成本保守重评估》。该存档包含复现与核验论文报告结果所需的复用基准数据集、提示词（Prompt）、实验输出、评估笔记本与配图。为遵循双盲评审要求，已移除包含作者身份的元数据与本地环境痕迹。