<h1 id="adeline-likefollow-8q2zx8qb8"><a rel="ugc" href="https://www.pixiv.net/users/1289777">Adeline</a> ⬅❤Like➕Follow</h1>I don’t understand why the effect gets worse after adding negative prompt words. (Is it possible that some negative things in the model are its style?)我不明白为什么在加入负面提示词后效果会变差。(难道模型中一些负面的东西就是它的风格？)

Files with SHA256: 216ec396cdfea73e2fd7c9259e29acfc216c0b3e163e800b7564ab45d3f41e8b