You're working with large datasets in Pandas API on Spark and notice performance issues with operations involving the default index. Which configuration option should you adjust to specify the default index type for better performance?