Fetching the paper…

DiaHalu: A Dialogue-level Hallucination Evaluation Benchmark for Large Language Models · Around