Open-Source LLM-Driven Formal Verification: A Multi-Agent Pipeline for RTL Repair

Researchers have introduced a multi-agent pipeline that utilizes open-source large language models to automate the repair of register-transfer level code. This approach aims to reduce the high costs and restrictive licensing associated with current formal verification tools used in chip design. By applying LLMs to verify hardware logic, the method seeks to lower the barriers to entry for mathematical correctness testing in semiconductor development.