Xuefei Lu - Unveiling the Path to Desired Predictions: An Interpretability Approach for Black-Box Models