Reproduce “Predicting Unseen Scientific Interactions: Conditional Flow Models Hallucinate Non-Additive Effects” 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
Sam Gijsen, Moritz Seiler, Kerstin Ritter
The authors assess conditional flow models’ recovery of two-factor, mean-based interactions in unseen scientific conditions. They distinguish aligned recovery and orthogonal hallucination from endpoint fidelity. In a controlled PPG simulator, matched held-out high–high and low–low corners show recovery of about 1.00 versus 0.31 and hallucination of about 0.08 versus 0.43. The scope excludes higher-order interactions and full distributional contrasts.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.