How to understand the high embedding vector value after training the model?

Dear PyTorch community,

Greetings, always thank you for your effort in the active discussion about the overall deep learning paradigm.
I have a general question about the embedding vector from the trained model.

I checked that my model returned node embedding vector after model training which results always predict samples into positive.

model = MyModel(in_channels=256, hidden_channels=128, out_channels=64).to(device) #if sum, in_channels=256
optimizer = torch.optim.Adam(model.parameters(), lr=1e-3, weight_decay=1e-3)
criterion = torch.nn.BCEWithLogitsLoss()

model.train()
for data in tqdm(train_loader):
    
    data = data.to(device)
    data.edge_class = data.edge_class[data.input_id]
    data.edge_index_class = torch.zeros(len(data.edge_index[0])).to(device)    
            
    for i in range(len(data.edge_index_class)):
        if ((data.edge_index[1][i] in data.edge_label_index[0]) or (data.edge_index[1][i] in data.edge_label_index[1])):
            data.edge_index_class[i] = data.edge_class[(data.edge_index[1][i]==data.edge_label_index[0])
                                                         |(data.edge_index[1][i]==data.edge_label_index[1])]
        else:
            data.edge_index_class[i] = torch.max(data.edge_index_class[(data.edge_index[1][i]==data.edge_index[0])])
            
    optimizer.zero_grad()
    
    z, a1, a2 = model(data.x[0], data.x[1], data.n_id, data.edge_index_class, data.edge_index) 
    
    out = ((z[data.edge_label_index[0]] * z[data.edge_label_index[1]]).sum(dim=-1)).view(-1) # product of a pair of nodes on each edge
    out_sig = ((z[data.edge_label_index[0]] * z[data.edge_label_index[1]]).sum(dim=-1)).view(-1).sigmoid()

print(out)
#tensor([ 8.4500,  7.3036,  7.7035,  8.7444, 11.7715,  8.8596,  7.1483, 11.5223,
#        11.2428, 10.7440,  7.6953, 10.4081,  7.4365,  7.4336,  9.0127,  7.7171,
#        10.3317,  8.9579,  6.8032, 11.4876,  7.3854,  5.2074,  6.3047, 10.6670,
#         7.2817,  9.7236,  6.9220, 10.3020], device='cuda:0',
#       grad_fn=<ViewBackward0>)

print(out_sig)
#tensor([0.9998, 0.9993, 0.9995, 0.9998, 1.0000, 0.9999, 0.9992, 1.0000, 1.0000,
#        1.0000, 0.9995, 1.0000, 0.9994, 0.9994, 0.9999, 0.9996, 1.0000, 0.9999,
#        0.9989, 1.0000, 0.9994, 0.9946, 0.9982, 1.0000, 0.9993, 0.9999, 0.9990,
#        1.0000], device='cuda:0', grad_fn=<SigmoidBackward0>)

print(data.edge_label)
# tensor([1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 1, 0,
#        1, 1, 1, 1], device='cuda:0')

In this case, where should I check the model?
The experimental setting is as follows:
dropout rate: 0.2
batch size: 32
learning rate: 0.001
weight decay: 0.001

Thank you for reading my question.