Reproduce “Polar: LLM-Powered Synthesis of Real-World Cyber Evidence for Prioritization and Mitigation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Luoxi Tang, Yuqiao Meng, Ankita Patra, Weicheng Ma, Muchao Ye, Zhaohan Xi
POLAR is an LLM framework that builds source-linked, time-ordered profiles of threats to prioritize vulnerabilities and recommend mitigations. It separates threats in reports, enriches profiles with evidence, assesses severity and exploitation likelihood over a 30-day horizon, and ranks responses. On threat-priority ranking, POLAR with GPT-5 achieves NDCG@5 of 87.3 ± 0.7%, versus 59.8 ± 1.2% for AUTOCVSS, on public-source evaluation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.