Reproduce “Beyond Risk Cues: A Mechanism-Level Benchmark and Verifiable Debate for Scam Video Detection on the Web” 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

Yifan Wang, Junyu Lu, Shun Zhang, Chaozhuo Li, Fanliang Bu, Rong Yin, Lingbin Bu, Changrong Min and 1 more

ScamCourt evaluates YouTube scam videos by comparing opposing arguments grounded in video evidence, while MechScam provides a human-verified benchmark of 6,614 videos spanning six scam mechanisms and six benign contexts. A checker verifies citation sources, and a trained judge assesses interpretations. On GLM-4.1V (9B), it reaches 94.4% accuracy versus 92.9% for same-backbone SFT, on this YouTube benchmark.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.