A system to rate how much an AI's advice can be trusted

Original: Epistemic Warrant for LLM Recommendations: Characterizing the Basis for Reliance When Ground Truth Is Unavailable

Source: arxiv.org ↗