12th Annual Best Scientific Cybersecurity Paper Competition Winners
The National Security Agency (NSA) has awarded the 12th Annual Best Scientific Cybersecurity Paper Competition to “Decoding Trust: Comprehensive Assessment of Trustworthiness in GPT Models.” The winning paper, authored by 19 researchers including Professors Dawn Song (UC Berkeley), Bo Li (UIUC), and Sanmi Koyejo (Stanford), proposes a groundbreaking evaluation framework for assessing the trustworthiness of large language models (LLMs) such as GPTs.
Dr. Adam Tagert, Technical Director of NSA’s Science of Security program, praised the work: “This framework is a noteworthy foundational advancement, providing essential metrics for evaluating and comparing future LLMs rigorously.”
Two papers received honorable mentions this year:
- “Ethical Frameworks and Computer Security Trolley Problems: Foundations for Conversations” by Tadayoshi Kohno (University of Washington), Yasemin Acar (Paderborn University/George Washington University), and Wulf Loh (University of Tübingen).
- “SoK: I Have the (Developer) Power! Sample Size Estimation for Statistical Tests in Developer-Centered Usable Security” by Anna-Marie Ortloff, Christian Tiefenau, and Matthew Smith (University of Bonn).
Nominations for the 13th Annual Competition, recognizing papers published in 2024, will open on January 15, 2025. Winners will be announced at the end of 2025.
Submitted by Regan Williams
on