Reproduce “Scaling Lean Theorem Proving for Real-World Code Verification” 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
Debargha Ganguly, Divya Raghunathan, Dejan Jovanović, Vikash Singh, Soonho Kong, Aaron Tomb, Aman Goel, Ferhat Erata and 1 more
EasyProver resolves 597/600 NTP4VC tasks, with 588 proofs and 9 refutations. It uses recursive agents and parallel tactic probing, with Lean’s kernel checking each result. Aegis generates properties for translated Rust implementations and freezes them as Lean statements. EasyProver proves 1,054 of a stratified 2,000-statement sample within 30-minute limits, showing these tasks remain challenging but not establishing Rust equivalence or specification fidelity.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.