Auto model validation

Yin

From what I have read auto model splits the data into 60:40% and then splits the 40 into 7 subsets and performs the scoring on those and produces the performance as average. Is the hyperparamter tuning or validation done using the 40%? If so how can this 40 then be used for scoring when the validation was already done on it?

I'm wondering if someone can explain automodel in terms of training, validation and testing/scoring. I tried reading the previous threads, but they don't describe this. I'm concerned if the validation set was the same as the set that the model used for scoring.

I really appreciate any feedback. 

PS: I am doing classification, and used all the valuable classifiers except "deep learning".

Thank you in advance.