1 comment

[ 2.5 ms ] story [ 12.1 ms ] thread
This is pretty cool to see. Glad to see the weights published. Currently it seems the best we have is OpenAI's `omni-moderation` model series or private models by companies like ActiveFence etc.