CoPE - the Content Policy Evaluation model

[samidh]

CoPE is an open model we specifically trained for policy-directed ("bring your own policy") content classification. It is optimized for fast inference while maintaining high accuracy. Details here: cope-a-9b huggingface repo

Eager, as always, for feedback on how this model performs on your use case.

[julietshen]

I love CoPE! It's small and fast, and I think the flexibility in its response format makes it a good choice for a lot of teams. I'm working on swapping it in for a small hobby project I'm playing with!