EI networks and Wilson Cowan model terms

What do the interaction weights between the excitatory and inhibitory subpopulations represent? Do they stand for some kind of averaged synaptic weight?