https://www.pixiv.net/users/91332959

这是模仿了哪位画师的风格 感觉质量挺高的

Base +6, funtaxi +6, xianfanneng +6, 空::白 +2, Zcmirch +4, si_hua +2, crsl +5, 天蝎星龙 +1, exyoyoyo +6, katiga -3

可以啊这年头AI无论在哪方面都相当出头，画风略像晰帅

Base +1, 空::白 +2, xexeeds +6, 天蝎星龙 +1, katiga -3

A lot of these seem pretty similar. Do you know which artists the model was trained on? Was the prompt keyed to any particular artist?

… I kinda wanna find the sources.

Base +6, Tentloei01 -6, lamprea12 +6, exyoyoyo +6, katiga -3

These are personally trained models and LoRA, it is impossible to trace the artists used, they become relevant precisely because they have unique models.
Today an AI model can completely change the source from which they start.

Base +6, albertzeb -6, Zeref -7, Verganza -6, Crockan +13, Tentloei01 -6, pencilfight -7, nonShade -13, pickaxeman -6, GefreiterKarl -8, weegeetoast -7, covs -6, katiga -3, and 2 more...

If that were true, than the AI model would have no trouble keeping its collages “in character.” This one seems to be referencing a much more limited scope than what you’re describing, as it’s struggling to use the same voice in each image. There also seems to be pockets of “this is this artist” and “this is that artist” that make me think these are specific individuals who are being referenced - or possibly specific pictures being used as a guide?

Base +6, albertzeb +6, Verganza +6, Tentloei01 -6, pencilfight +7, nonShade +13, katiga -3, greyneptune -6

No, it's not as simple as you describe, at the moment many anime-inspired prompters are using Agemo AGM Style a LYCORIS, but if he is personally trained can mix it with hundreds of other similar artists (or prompters!! Do you know that you can train your AI on results from other prompters? Again, do you know that you can train your AI on the results of YOUR models in an infinite mix?).
And not only that, he can retouch them on the output and re-insert them into the model, the inconsistencies of an AI will always be there, especially if you have mixed a lot of them.
In this gallery it is evident that he is using a lot of LoRA for style, face, positions...But the prompt triggers for cohesion are trained by him.

Tldr: Personally trained models and LoRA, it is impossible with certainty to trace the artists used.

Base +6, albertzeb -6, Zeref -7, Verganza -6, Crockan +13, Tentloei01 -6, pencilfight -7, pickaxeman -6, a84 -6, GefreiterKarl -8, yang0 +6, covs -6, katiga -3, and 3 more...

这些家伙都是不肯放原图的啊

Base +6, katiga -3

大部分都是杂图，能不能正戏多点，都 AI 了我还看啥美图啊？我需要更多性交！！

Base +5, 云上花开 +6, katiga -3, evej4360 +7, wujl +6, 想草伊莉雅 +2, 魈Xiao61991 +5

好像知道他们是用的什么软件做的

Base +1, katiga -3