The doubt is that if activation means degree of high output then how it can be output for other neuron what it will do by knowing the degree of output of neuron ?
A high activation value means the related feature is important in predicting the output of the model for that example.