AI & ML

Guardrails

The technical and policy measures that constrain a model’s outputs — content filters, system prompts, and refusal training among them.