XGBoost- default values and start

Hi, i am new to XGBoost. I am trying to understand that at the beginning of the XGBoost regressor, is it correct to state that, “the first step of the XGBoost model is to create a hypothesis to fit it into the training dataset by making an initial prediction of residuals, and the initial prediction is set to the mean of targeted training dataset values.” Also, in some blogs, I have seen that it is set to 0.5.
Can someone give me reliable documentation that explains how it is taken? Also, the default learning rate is 0.3, but in my code it is set as none.