Let owners choose what should happen to messages with scores over particular thresholds. For example, "automatically hold messages with toxicity > 0.8 for moderator review"
we can reuse a lot of existing tagging related code if the setting is implemented as: if score above 0.8, apply tag "toxic". and then let the user decide what happens to messages with tag toxic.
Let owners choose what should happen to messages with scores over particular thresholds. For example, "automatically hold messages with toxicity > 0.8 for moderator review"