University of Maryland — LLM Cybersecurity Benchmark

[Navigating Transformative AI] Open Philanthropy recommended a grant of $1,721,079 over two years to the University of Maryland to support work led by Professor Yizheng Chen to develop a benchmark to assess the cybersecurity capabilities of large language models. This grant was funded via a request for proposals for projects benchmarking LLM agents on consequential real-world tasks. This falls within our focus area of potential risks from advanced artificial intelligence.