Exploring Musical Roots: Applying Audio Embeddings to Empower Influence Attribution for a Generative Music Model