Yansheng Qiu, Haoquan Zhang, Zhaopan Xu, Ming Li, Diping Song, Zheng Wang, Kaipeng Zhang
We present AI Idea Bench 2025, a framework to quantitatively evaluate LLM-generated ideas in AI research, featuring a dataset of 3,495 papers with inspired works and a robust evaluation methodology. The system assesses idea quality based on alignment with original papers and reference-based judgment.
Paper Project Page Code Benchmark