Improving Factuality in LLMs via Inference-Time Knowledge Graph Construction
This addresses the issue of factual errors in LLMs for users relying on accurate information, representing an incremental improvement over existing retrieval-augmented methods.
The paper tackles the problem of factual inconsistency in LLM outputs by proposing a framework that dynamically constructs and expands knowledge graphs during inference, integrating internal and external knowledge, and demonstrates consistent gains in factual accuracy on three Factual QA benchmarks.
Large Language Models (LLMs) often struggle with producing factually consistent answers due to limitations in their parametric memory. Retrieval-Augmented Generation (RAG) paradigms mitigate this issue by incorporating external knowledge at inference time. However, such methods typically handle knowledge as unstructured text, which reduces retrieval accuracy, hinders compositional reasoning, and amplifies the influence of irrelevant information on the factual consistency of LLM outputs. To overcome these limitations, we propose a novel framework that dynamically constructs and expands knowledge graphs (KGs) during inference, integrating both internal knowledge extracted from LLMs and external knowledge retrieved from external sources. Our method begins by extracting a seed KG from the question via prompting, followed by iterative expansion using the LLM's internal knowledge. The KG is then selectively refined through external retrieval, enhancing factual coverage and correcting inaccuracies. We evaluate our approach on three diverse Factual QA benchmarks, demonstrating consistent gains in factual accuracy over baselines. Our findings reveal that inference-time KG construction is a promising direction for enhancing LLM factuality in a structured, interpretable, and scalable manner.