Reproduce “Atlas: Making Sense of Long Horizon Agent Trajectories” 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
Sangho Suh, Joseph Chee Chang, Pao Siangliulue, Amy X. Zhang, Matt Latzke, Jonathan Bragg, Doug Downey, Daniel S. Weld
Atlas helps researchers review completed research-agent runs by linking the research goal, the agent’s decisions, and its outputs in a branching trajectory and artifact views. In a within-subjects study of 14 researchers, decision-transparency ratings were higher with Atlas than with the standard workspace: median 6.0 versus 3.8 on a 7-point scale. These self-reports do not establish research accuracy or quality.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.