Diagnostic Reasoning Prompts Reveal the Potential for Large Language Model Interpretability in Medicine