GPTKB v1.5: A Massive Knowledge Base for Exploring Factual LLM Knowledge

TL;DR

This paper introduces GPTKB v1.5, a large-scale, densely interlinked knowledge base built from GPT-4.1, enabling improved exploration, querying, and analysis of factual knowledge in language models.

Contribution

It presents a novel methodology for massive-recursive LLM knowledge materialization and demonstrates its application through a comprehensive knowledge base and interactive exploration tools.

Findings

Knowledge base contains 100 million triples

Enables link-traversal and SPARQL-based querying

Facilitates systematic analysis of LLM factual knowledge

Abstract

Language models are powerful tools, yet their factual knowledge is still poorly understood, and inaccessible to ad-hoc browsing and scalable statistical analysis. This demonstration introduces GPTKB v1.5, a densely interlinked 100-million-triple knowledge base (KB) built for $14,000 from GPT-4.1, using the GPTKB methodology for massive-recursive LLM knowledge materialization (Hu et al., ACL 2025). The demonstration experience focuses on three use cases: (1) link-traversal-based LLM knowledge exploration, (2) SPARQL-based structured LLM knowledge querying, (3) comparative exploration of the strengths and weaknesses of LLM knowledge. Massive-recursive LLM knowledge materialization is a groundbreaking opportunity both for the research area of systematic analysis of LLM knowledge, as well as for automated KB construction. The GPTKB demonstrator is accessible at https://gptkb.org.