I'm a big fan of merging, but specifically when the two characters mouths are connected. This model is to try and replicate that part specifcally but it also works for merging in general.
Main trigger words are (conjoined mouth) and (merging).

For best results of getting the mouth connected, use (facing viewer) and (face focus) or (mouth focus). Higher step count also helps.

Works better with canine characters, but can work with other face shapes as well.

Will usually make the characters fat when making a full body image, so include a prompt for body shape if this isn't wanted.

Increasing strength will help, but will also influence style