Why not create a llm model to classify low quality posts and test by posting to future low quality posts. If that works, use the reddit api to moderate based on model predictions?

It beats time spent frustrating yourself looking through posts you don’t want to see.