RAG-Verus: Repository-Level Program Verification with LLMs using Retrieval Augmented Generation
Zhong, Sicheng, Zhu, Jiading, Tian, Yifang, Si, Xujie
–arXiv.org Artificial Intelligence
Scaling automated formal verification to real-world projects requires resolving cross-module dependencies and global contexts, which are challenges overlooked by existing function-centric methods. We introduce RagVerus, a framework that synergizes retrieval-augmented generation with context-aware prompting to automate proof synthesis for multi-module repositories, achieving a 27% relative improvement on our novel RepoVBench benchmark -- the first repository-level dataset for Verus with 383 proof completion tasks. RagVerus triples proof pass rates on existing benchmarks under constrained language model budgets, demonstrating a scalable and sample-efficient verification.
arXiv.org Artificial Intelligence
Feb-7-2025
- Country:
- North America
- United States > California
- Los Angeles County > Los Angeles (0.14)
- Canada > Ontario
- Toronto (0.14)
- United States > California
- North America
- Genre:
- Research Report (0.50)
- Technology: