Reproduce “Type-Safe Is Not Error-Free: Typed Decision Models Follow the Option Name, Not the Definition Bound to It” 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
Yu Sun, Junhao Xu, Jiajia Shi, Zijin Yang
The authors test whether typed decision models follow option definitions or familiar names by swapping the definitions bound to fixed names while keeping questions and choices unchanged. On 1,200 questions, Laya’s definition-level flip rate is 76.92% for yes/no and 6.50% for 0/1. Yes/no AUC falls from 93.8% to 23.2%; outputs remain type-correct, limiting evidence to option-name polarity’s contribution to instability.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.