Interpreting CT-Scans with CLIP: An Explorative Study of Attribution Methods for 3D Vision-Language Models