Context-Dependent Personalised Feedback Prioritisation in Exploratory Learning for Mathematical Generalisation