A focused specialist rather than a generalist — this model exists specifically to evaluate whether content is safe or harmful, not to generate creative text or answer questions. It operates as a classifier and filter, making binary or categorical judgments about inputs. Its narrow purpose means it excels at what it does but isn't designed for open-ended conversation.