Pymc5 out of sample

Hi Jesse! thank you.
Now the code is running correctly, and I can perform an out-of-sample run with this code.

with basic_model_2:
    pm.set_data({"feature_data":x_test, "control_data":z_test})
    idata.extend(pm.sample_posterior_predictive(idata))
    
y_hat_predicted = idata.predictions['y_hat'].mean(dim=["chain", "draw"])

However, the ‘idata’ object in ‘predictions’ level only contains the estimated value of y_hat.
It would be fantastic to have the values of the individual elements that compose it.
In other words, the prediction of each feature and each control.

Thank you!