It depends on what you mean by "models". LLMs? No. But LLMs are too slow for content moderation at scale. Custom trained models? Maybe. Is the unicode charact

It depends on what you mean by "models". LLMs? No. But LLMs are too slow for content moderation at scale. Custom trained models? Maybe. Is the unicode characters in the training data?