Graph-Grounded LLMs: Leveraging Graphical Function Calling to Minimize LLM Hallucinations
This addresses the issue of unreliable LLM outputs in graph-based applications like autonomous vehicles and knowledge graphs, though it is an incremental improvement by adding external tools.
The paper tackles the problem of LLM hallucinations and inaccuracies in graph-related tasks by proposing Graph-Grounded LLMs, which integrate a graph library via function calls, resulting in significant reductions in hallucinations and improved mathematical accuracy as shown on the NLGraph benchmark.
The adoption of Large Language Models (LLMs) is rapidly expanding across various tasks that involve inherent graphical structures. Graphs are integral to a wide range of applications, including motion planning for autonomous vehicles, social networks, scene understanding, and knowledge graphs. Many problems, even those not initially perceived as graph-based, can be effectively addressed through graph theory. However, when applied to these tasks, LLMs often encounter challenges, such as hallucinations and mathematical inaccuracies. To overcome these limitations, we propose Graph-Grounded LLMs, a system that improves LLM performance on graph-related tasks by integrating a graph library through function calls. By grounding LLMs in this manner, we demonstrate significant reductions in hallucinations and improved mathematical accuracy in solving graph-based problems, as evidenced by the performance on the NLGraph benchmark. Finally, we showcase a disaster rescue application where the Graph-Grounded LLM acts as a decision-support system.