You are using Vertex AI and TensorFlow to develop a custom image classification model for your company. Transparency and interpretability of the model’s decisions are critical to ensure trust and buy-in from stakeholders. Additionally, you need to explore the results to identify any issues or potential biases that may exist within the model predictions. What should you do to achieve these goals?