Infowars / Zohran

Not perfect, but not bad either. The model reliably tagged deterministic, inevitability-heavy claims, broad stereotyping, and escalating agitation — the patterns that can mislead by substituting certainty and caricature for evidence. It also correctly elevated the most consequential handling risk: serious allegations presented without clear verification or right of reply.