Post #705526

LLM-Driven Discovery of High-Entropy Catalysts via Retrieval-Augmented Generation CO2 reduction requires efficient catalysts, yet materials discovery remains bottlenecked by 10-20 year development cycles requiring deep domain expertise. This paper demonstrates how large language models can assist the catalyst discovery process Data Access: Downloaded from Open Catalyst Project repositories as PyTorch Geometric Data objects and #LMDB files. https://arxiv.org/html/2603.15712v1