When to Trust Context: Self-Reflective Debates for Context Reliability

Zeqi Zhou, Fang Wu, Shayan Talaei +5

Large language models frequently encounter conflicts between their parametric knowledge and contextual input, often resulting in factual inconsistencies or hallucinations. We propo…