{"ID":2846079,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.02285","arxiv_id":"2511.02285","title":"VFocus: Better Verilog Generation from Large Language Model via Focused Reasoning","abstract":"Large Language Models (LLMs) have shown impressive potential in generating Verilog codes, but ensuring functional correctness remains a challenge. Existing approaches often rely on self-consistency or simulation feedback to select the best candidate, but they miss opportunities to focus LLM reasoning on the most informative parts of the design. We propose VFocus, a three-stage framework that enhances Verilog generation by sharpening the focus of LLM reasoning onto critical decision points in the code generation process. In the \\textbf{pre-ranking stage}, VFocus generates multiple code candidates through LLM prompting, retries for syntactically valid outputs, and introduces a \\textit{Density-guided Filtering} to retain candidates that fall within the \"reasoning sweet spot\" for functional correctness. In the \\textbf{ranking stage}, we simulate each code candidate using an automatically generated testbench and apply self-consistency-based clustering to identify the most consistent outputs. Finally, in the \\textbf{post-ranking refinement stage}, VFocus performs inconsistency mining on top-ranked candidates and invokes reasoning-augmented LLM prompts for candidate refinement. Experiments on the VerilogEval-Human benchmark show that VFocus significantly improves the pass@1 correctness across multiple reasoning LLMs, demonstrating its effectiveness in enhancing Verilog generation for complex hardware design tasks.","short_abstract":"Large Language Models (LLMs) have shown impressive potential in generating Verilog codes, but ensuring functional correctness remains a challenge. Existing approaches often rely on self-consistency or simulation feedback to select the best candidate, but they miss opportunities to focus LLM reasoning on the most inform...","url_abs":"https://arxiv.org/abs/2511.02285","url_pdf":"https://arxiv.org/pdf/2511.02285v1","authors":"[\"Zhuorui Zhao\",\"Bing Li\",\"Grace Li Zhang\",\"Ulf Schlichtmann\"]","published":"2025-11-04T05:54:31Z","proceeding":"cs.AR","tasks":"[\"cs.AR\",\"cs.PL\",\"cs.SE\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false}