Get Started
Research questionHow can LLM-built knowledge bases prevent duplicate or conflated entities, relations, and classes at scale?LLMs lack stable native representations of entities, so the same entity may produce duplicate entries or distinct entities may be merged. Large-scale construction must resolve these ambiguities across entities, relations, and classes without making extraction prohibitively costly.
AI
Information Retrieval
Natural Language Processing
Latest papersRecent research connected to this question, newest first.Direct Construction of Disambiguated Knowledge Bases from Large Language ModelsThe source describes GPTKB 2.0, which performs on-the-fly disambiguation and explicit internal canonicalization while materializing more than 1 million entities and 38.4 million triples. Its evidence addresses trade-offs among accuracy, scale, and cost for direct LLM-derived knowledge-base construction.research paper · Sep 2, 2026
Related questions
How can LLM-derived knowledge bases disambiguate homonymous entities and preserve auditable fact provenance?How can we construct accurate knowledge bases from an LLM’s parametric knowledge without retrieval, fine-tuning, or exceeding 32B parameters?How can LLMs remain faithful to provided context without doubling inference cost?How can we evaluate LLM knowledge updates over time without contamination or inconsistent facts?
Home
Topics
Search
Library