Reproduce “Evidence-Guided Repository-Level RTL Repair” 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
Yuxin Du, Juxin Niu, Zhe Jiang, Nan Guan
The framework repairs repository-level RTL bugs by reproducing a failure, tracing its waveform with structured queries, and propagating candidate edits across connected design sites. It replays the original scenario to check whether the violation is resolved. With DeepSeek V4 Flash and the same OpenHands SDK scaffold, it resolves 312 of 388 tasks, versus 273; results exclude 29 unadapted OpenTitan tasks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.