Reasoning: Med-V1: Small Language Models for Zero-shot and Scalable Biomedical Evidence Attribution | Foresight | ScienceToStartup