Cellular signaling networks represent complex information processing systems that have been modeled via traditional mathematical or statistical approaches. However, these methods often struggle to capture context-dependent signaling, pathway cross-talk, and temporal dynamics across multiple biological scales. Here, we introduce hierarchical molecular language models (HMLMs), a novel architecture that proposes a molecular network-specific large language model (LLM) to use in intracellular communi...