This LoRA is created to generate pictures of men groping women who are in a vulnerable state of dress (like in their underwear.) It's something SD1.5 hasn't been able to do very well out of the box so I generated some images of this concept using Flux and trained a SD1.5 LoRA on them. If you try to prompt for a man and woman separately and wearing different clothing in most SD1.5 models, you'll either get two men, two women, and/or a man and woman all wearing similar clothing or of the same ethnicity. This LoRA hopefully decouples them without needing fancy addons in A111 like Latent Couple.
This LoRA doesn't really have any trigger words but you should prompt it in a format similar to the following:
man groping blindfolded woman in her bra and panties
Note: here are what sorts of attire I've trained the man and the woman to have, though you may need to have the groping keyword in your prompt and define how the man and the woman should look in either side of the keyword.
Man:
shirtless man
man in soldier uniform
Woman:
bra
t-shirt
Japanese school shirt
Another example: shirtless man groping blindfolded woman in Japanese school shirt and panties looking at viewer.
It may work with different styles of clothing or without blindfolds. Also, you may need to specify the ethnicity of the people if you wish to go with something other than the defaults.
Try it with animations!