Hi,
For this purpose, you can use the bundled Gradient Probability macro, where you can fine-tune the Transition Position and Transition Range parameters to control the distribution. Please find attached a modified version of your scene. I hope this helps.
Best regards,