Hallucination

Large language models are prone to generating hallucinations, such as factual errors and dedicated falsehood.