The availability of large-scale neuronal population datasets necessitates new methods to model population dynamics and extract interpretable, scientifically translatable insights. While deep learning offers potential solutions, existing models often overlook the biological mechanisms underlying population activity and thus exhibit suboptimal performance with neural data and provide little to no interpretable information about neurons and their interactions. In response, we introduce \model{}, a novel deep-learning framework that effectively models population dynamics and the interactions between neurons. Within this biologically realistic framework, each neuron, characterized by a latent embedding, sends and receives currents through directed connections. A shared decoder uses the input current, previous neuronal activity, neuron embedding, and behavioral data to predict the population activity in the next time step. Our experiments demonstrate that \model{} consistently outperforms existing models in forecasting population activity. Additionally, our experiments on synthetic data showed that \model{} accurately recovers ground-truth connections between neurons.