用知识图谱向量增强大模型，不改结构、少用token，事实记忆提升超2倍。

Retrieval Augmented Generation (RAG) has enjoyed increased attention in the recent past and recent advancements in Large Language Models (LLMs) have highlighted the importance of integrating world knowledge into these systems. Current RAG methodologies often modify the internal architecture of pre-trained language models (PLMs) or rely on textifying knowledge graphs (KGs), which is inefficient in terms of token usage. This paper introduces ConceptFormer, a new approach to augment LLMs with structured knowledge from KGs, such as Wikidata, without altering their internal structure or relying on textual input of KGs. ConceptFormer operates in the LLM embedding vector space, creating and injecting \emph{concept vectors} that encapsulate the information of the KG nodes directly. Trained in conjunction with a frozen LLM, ConceptFormer generates a comprehensive lookup table that maps KG nodes to their respective concept vectors. The approach aims to enhance the factual recall capabilities of LLMs by enabling them to process these concept vectors natively, thus enriching them with structured world knowledge in an efficient and scalable manner. Our experiments demonstrate that the addition of concept vectors to GPT-2 0.1B substantially increases its factual recall ability (Hit@10) by up to 272\% when tested on sentences from Wikipedia and up to 348\% on synthetically generated sentences. Even injecting only a single concept vector into the prompt increases factual recall ability (Hit@10) by up to 213\% on Wikipedia sentences, significantly outperforming RAG with graph textification while consuming 130x fewer input tokens.