Towards Neuron Attributions in Multi-Modal Large Language Models