What do the results by Model-to-Model Distance means?

I get a problem about the meaning of value from Model to Model Distance. I set the predicted model as source model and the ground truth model as target model, I thought that the negative value means the predicted model is inside the ground truth model. But I got the same result when I switched them (the predicted model as target model and the ground truth model as source model)

I am very confuse. What is wrong?

I get a problem about the meaning of value from Model to Model Distance. I set the predicted model as source model and the ground truth model as target model, I thought that the negative value means the ground truth model is inside the predicted model.

But the another result, I got the same result when I switched them (the predicted model as target model and the ground truth model as source model)

I am very confuse. What is matter?