One option would be to insert a low probability clip randomly in a set of three trials and then follow this set with a high probability clip before starting again.
Have a loop at Complex Randomisation and Randomisation without repetition as two approaches in my PsychoPu online demos