핵심 개념
Enhancing attribution in large language models is crucial for accurate citations and grounded summaries.
통계
WebCiteS dataset features 7k human-annotated summaries with citations.
WebGLM prompts InstructGPT to generate training data with citations.
ALCE benchmark compiles question-answering datasets without citations.
인용구
"Enhancing the attribution in large language models is a crucial task."
"Our comprehensive evaluation highlights the challenge LLMs face in correctly citing sources."