Building an early warning system for LLM-aided biological threat creation
We’re developing a blueprint for evaluating the risk that a large language model (LLM) could aid someone in creating a biological threat.