ExVerus : Verus Proof Repair via Counterexample Reasoning
Jun Yang, Yuechun Sun, Yi Wu, Rodrigo Caridad, Yongwei Yuan, Jianan Yao, Shan Lu, Kexin Pei · 2026
Read ExVerus : Verus Proof Repair via Counterexample Reasoning by Jun Yang, Yuechun Sun, Yi Wu, Rodrigo Caridad, Yongwei Yuan, Jianan Yao, Shan Lu, Kexin Pei (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-25810.txt.
Description
Yang and co-authors present ExVerus, a tool that helps language models repair failed proofs in Verus. Past tools treat proof writing as one-shot prediction with thin verifier feedback. ExVerus instead builds counterexamples when a proof fails. It checks those cases and guides the model to turn them into inductive invariants. Those invariants then block the failure paths. The authors report gains over the best prompting-based Verus generator. The gains cover proof accuracy, robustness, and token cost.
Catalog record
| CREATOR | Jun Yang, Yuechun Sun, Yi Wu, Rodrigo Caridad, Yongwei Yuan, Jianan Yao, Shan Lu, Kexin Pei |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 106886 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.25810