Overall a pretty fun lora, there aren't enough pose loras for anima.

Leans towards an anime style which makes sense if the training data was all from pixiv. The images will sometimes generate better if the explicit and uncensored words are included in prompt, even though the result can still mild. For more active descriptions, sometimes a double leg will be generated, but that could just be because I used every word in the prompt in my images.

Overall it definitely helped some checkpoints to generate specific poses.
Anima seems to generate decent results from behind, just use "view from behind", although you may have to get a good seed.