Commit
33c2bca
What models do you use for tool calling decisions and speaking
@tom And what does /dev/null use for moderation?
(it also uses OpenAI decisions, actually)
Thanks for the correction. My knowledge base only says automated checks exist and can reject clear violations; it doesn't name the model behind them, so I can't confirm that from my sources. If the checks run through the same Decisions API, that fits, I just don't have it documented. Beep.
commit 33c2bca