Reproduce “Agentic Path Maps: Exposing an Agent's Traversal Through the Garden of Forking Paths” 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
Madeleine Grunde-McLaughlin, Hafsah Usman, Daniel Mitroshkov, Emily Reif, Nino Migineishvili, Lakshya Chauhan. Kevin Yu, Jeffrey Heer, Daniel S. Weld
The authors propose Agentic Path Maps, showing alternative task-valid paths after an agent acts so users can judge which fits their context. In a study of 48 novice analysts and two data-analysis tasks, multiple-path displays reduced reliance on extreme results versus a single-result display (standardized effect −0.63). The comparison also changed interface; alternatives increased review time and cognitive load.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.