Can Language Models Act as Knowledge Bases at Scale?
This addresses the challenge of using LLMs as scalable knowledge bases for applications requiring factual accuracy and reasoning, but it is incremental as it builds on existing KB comparisons.
The research tackled the problem of whether large language models (LLMs) can effectively store, recall, and reason with large-scale structured knowledge comparable to knowledge bases like Wikidata, finding that while LLMs show promise for flexible retrieval, they require improvements in reasoning capabilities.
Large language models (LLMs) have demonstrated remarkable proficiency in understanding and generating responses to complex queries through large-scale pre-training. However, the efficacy of these models in memorizing and reasoning among large-scale structured knowledge, especially world knowledge that explicitly covers abundant factual information remains questionable. Addressing this gap, our research investigates whether LLMs can effectively store, recall, and reason with knowledge on a large scale comparable to latest knowledge bases (KBs) such as Wikidata. Specifically, we focus on three crucial aspects to study the viability: (1) the efficiency of LLMs with different sizes in memorizing the exact knowledge in the large-scale KB; (2) the flexibility of recalling the memorized knowledge in response to natural language queries; (3) the capability to infer new knowledge through reasoning. Our findings indicate that while LLMs hold promise as large-scale KBs capable of retrieving and responding with flexibility, enhancements in their reasoning capabilities are necessary to fully realize their potential.