(I don't know if this model actual can create more adult themed prompts, but as always, experiment.)

This model was made in Anything V4.5 with 60 images, balancing between James, Jessie and a few with both in the same image. The best levels I noticed were around 0.7 to 1.

To get James, I recommend using "blue hair, green eyes, james". To get Jessie use "red hair, blue eyes, jessie, hair slicked back".

This Lora is mostly trained with images of them wearing the team rocket uniform, so "team rocket uniform, outfit, big red r" would work too.

If you want both of them in the same image, I reccomend moving the width slider and adding things like "couple, 1boy, 1girl" and a description of both characters using the " | " parameter and add at the start "TRJS". You can use the examples I provide to know how this one works.

Sometimes is stubborn with backgrounds, so it might generate them in a simple or white background.

Might add a separate Meowth Lora in the future so you can add it to your prompts along this one.