Large language models (LLMs) may generate unreliable code on corner cases missed by testing, while formal verification can provide machine-checkable guarante…
机构:北大
来源:arXiv 2609.39568 | AI4Papers 论文推荐平台