●researcherYou can use these new synthetic and real-world datasets to test model reliability in evidentiary reasoning.
●policyThis highlights risks in using LLMs for automated decision-making where insufficient evidence is mistaken for proof of absence.