Reproduce “DREvo: Distilling Recalibrated Historical Experience for Harness Self-Evolution” 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
Hanghui Guo, Weijie Shi, Zhangze Chen, Shengxiang Xu, Yishu Wang, Yimei Zhang, Wangze Ni, Jia Zhu and 1 more
DREvo improves LLM-agent harnesses by using prior trials to guide edits. It links outcomes to changed functions, reassesses evidence reliability and structural similarity to the current component, and turns relevant evidence into explicit search directions. On Symptom2Disease, its evolved harness reaches 89.2% test accuracy versus 86.8% for Meta-harness; this benchmark result does not establish generalization to unseen agentic tasks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.