If you use

bucket_size: 2

for example, you'll get the first bucket with the two documents matching the score, and then sorted based on popularityd