DoD to develop scalable genAI testing datasets

Through a recently completed multi-pronged red teaming effort, the agency said it will develop repeatable testing datasets that can be used to evaluate large language model tools and services in the future.