This made for a very interesting test. I've posted a full set of images that use the same seed/settings alongside various prompts focusing on mainly lighting/reflection tags to see how it handles them.

My posted images should show the results of a prompt without the detailer, then one with it at 1.0 weight to compare. I didn't reroll any of the generations using the detailer taking the first image it generated as a true test, so one or two do have the usual Illustrious errors (the samurai without her sword in her hand for example)

Apart from one or two surprising changes that were probably the result of my minimal character tags alongside the variations the checkpoint brings (the cyborg & elemental looking drastically different for example, Cloud's gender change was absolutely my fault for not tagging a gender though :P) they were all certainly improved upon with better effects, poses and more dynamic angles all around.

I plan to do another round of tests with more lighting/shadow effects and maybe some weight adjustments to see how it reacts but i really like it overall! (b ᵔ▽ᵔ)b