A financial institution uses a foundation model to support loan decisions and requires the model’s decisions to be explainable for audit and security reasons. Which factor most directly affects the explainability of the model’s decisions?

Model complexity directly affects explainability because simpler models (e.g., linear regression, decision trees) are inherently easier to understand and interpret. Their decision-making processes are transparent, making it straightforward to trace how inputs lead to outputs. Conversely, highly complex models like deep neural networks often operate as "black boxes," making it difficult to discern the exact reasons behind a specific prediction, which is crucial for audit and security in financial institutions. Training time and deployment time are operational metrics that do not directly impact how understandable a model's internal logic is. The number of hyperparameters can contribute to complexity but is not the primary determinant of explainability itself; a model with many hyperparameters could still be simple in its architecture.