An organization uses MLflow for model tracking. To ensure auditability of model lineage, what must be captured in the 'mlflow.log_artifact' call?