Reproduce “SeqFeed: Improving Agentic RTL Code Generation with Sequential Behavior Feedback” 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, Tao Hu, Xi Wang, Zhe Jiang, Nan Guan
SeqFeed helps an RTL-debugging agent investigate a data-stream accumulator that returns 20 instead of the expected 15 by querying signal values around an accepted input with SeQuery and tracing cross-cycle dependencies with SeGraph. In one model setting, using both tools raises pass rate from 60.4% to 69.1% across 256 cases; this result does not establish performance for other settings.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.