This paper introduces a large-scale evaluation framework to measure the gap between LLM-generated research