Reproduce “Typed Decision Models on Tabular Data” 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
Baowen Zhang, Hangting Ye, Hongyuan Zha
The authors evaluate typed decision models for fixed-choice tabular tasks, returning probabilities over answer options rather than generated text. They compare them with large language models on classification, verification, and anomaly detection. On row classification, GPT-6 Luna Decisions scores 61.2 Macro-F1 versus 62.5 for GPT-6 Luna. Under at least one information-preserving table rewrite, decisions change for all tested models, leaving robustness unresolved.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.