SoCureLLM: An LLM-driven Approach for Large-Scale System-on-Chip Security Verification and Policy Generation

Authors

Shams Tarek, Dipayan Saha, Sujan Kumar Saha, Mark Tehranipoor, Farimah Farahmandi

Abstract

Contemporary methods for hardware security verification struggle with adaptability, scalability, and availability due to the increasing complexity of modern system-on-chips (SoCs). Large language models (LLMs) have emerged as a viable approach to address these shortcomings. This paper introduces SoCureLLM, an LLM-based framework designed to excel in identifying security vulnerabilities within SoC designs and creating a comprehensive security policy database. In evaluations, SoCureLLM detected 76.47% of security bugs across three vulnerable RISC-V SoCs, outperforming state-of-the-art security verification methods. Furthermore, assessing three additional large-scale RISC-V SoC designs against various threat models led to the formulation of 84 novel security policies.

@inproceedings{tarek2025socurellm,
  title={Socurellm: An llm-driven approach for large-scale system-on-chip security verification and policy generation},
  author={Tarek, Shams and Saha, Dipayan and Saha, Sujan Kumar and Tehranipoor, Mark and Farahmandi, Farimah},
  booktitle={2025 IEEE International Symposium on Hardware Oriented Security and Trust (HOST)},
  year={2025},
  organization={IEEE}
}