Reproduce “Decision-Centric Research: A Calibrated Decision Model and Benchmark for Scientific Agents” 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
Shaoning Zeng, Wen Huang, Xinzhong Zhu, Yi Duan, Jianhang Zhou, Jianping Gou
The authors introduce a state-aware decision layer that frames scientific choices as typed questions with defined candidate answers. It scores those candidates in one pass and provides probabilities for inspection and escalation. On RD-Bench, a 0.6B model reaches 89.4% Top-1 accuracy, versus 27.8% without state context. The benchmark measures agreement with rule-generated answers, not expert-validated research quality or end-to-end outcomes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.