Reproduce “Visible Reasoning: User-Facing Decision Transparency for Generative AI Systems” 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
Sanjay Krishna Anbalagan, Xinrui Nie, Anughna Kommalapati, Vijay Kumar Kanamarlapudi, Sreekanth Radhakrishnan, Xiaodan Zhao, Umesh Mohan
Filtered Decision Logs (FDLs) help users inspect recommendations from tool-using AI by presenting a concise rationale, ranked tool evidence, and controls to exclude a source and rerun. In three mock scenarios, removing the top-ranked source changed answers more than removing the bottom-ranked source; in the career case, similarities were 0.72 and 0.91. This signals sensitivity, not causal attribution or correctness.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.