Reproduce “Measuring AI "Slop" in Text” 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

Chantal Shaib, Tuhin Chakrabarty, Diego Garcia-Olano, Byron C. Wallace

The authors develop a taxonomy for judging “slop” in text, separate from whether it was AI-written. They organize expert definitions into seven labels and test how span annotations relate to holistic judgments across news and question-answering passages. All seven labels predict judgments, with different labels associated with news and QA. Syntax-only features reach AUROC 0.64, but automated detection remains modest.

Summary by alphaXiv.

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