Help center · Citations & sources

Do you track sarcasm or subtle nuance in AI responses?

Not reliably — sarcasm and heavy nuance are hard for automated sentiment analysis in general, and AI-generated answers tend to be fairly literal anyway, so this comes up rarely in practice. When a response reads as ambiguous, we lean neutral rather than guess at a stronger sentiment.