This is a Stable Diffusion 1.5 textual inversion embedding, <neg-anime>, trained using InvokeAI on a set of low quality generated sketches and pathological outputs from an animated model (https://civitai.com/models/35893/526mix-animated). Its use in a negative prompt helps push that model and others toward a more "realistic" style, sometimes. This was inspired by and heavily patterned after another embedding found here: https://huggingface.co/JPPhoto/neg-sketch-2