analiza aiqbench
This research introduces a novel framework for detecting LLM hallucinations by modeling the model itself as a dynamical system, leveraging Koopman operator theory to analyze latent state-space dynamics. By calculating differential residual scores from single-sample passes, the method achieves state-of-the-art accuracy without the computational overhead of secondary sampling or external knowledge retrieval. This approach significantly lowers the resource barrier for real-time hallucination monitoring in production environments.
Low-Cost Black-Box Detection of LLM Hallucinations via Dynamical System Prediction
arXiv:2605.05134v2 Announce Type: replace Abstract: Large Language Models (LLMs) frequently generate plausible but non-factual content, a phenomenon known as hallucination. While existing detection methods typically rely on computationally expensive sampling-based consistency checks or external knowledge retrieval, we propose a new method that treats the LLM as a black-box dynamical system. By projecting LLM responses into a high-dimensional manifold via an embedding model, we characterize the resulting vector sequences as observable realizations of the model's latent state-space dynamics. Leveraging Koopman operator theory, we fit the transition operators for both factual and hallucinated regimes and define a differential residual score based on their respective prediction errors. This approach enables low-cost hallucination detection in a single-sample pass, avoiding the need for secondary sampling or external grounding. Extensive testing across three data benchmarks demonstrates that our method achieves state-of-the-art performance with reduced resource overhead.
Abstract:Large Language Models (LLMs) frequently generate plausible but non-factual content, a phenomenon known as hallucination. While existing detection methods typically rely on computationally expensive sampling-based consistency checks or external knowledge retrieval, we propose a new method that treats the LLM as a black-box dynamical system. By projecting LLM responses into a high-dimensional manifold via an embedding model, we characterize the resulting vector sequences as observable realizations of the model's latent state-space dynamics. Leveraging Koopman operator theory, we fit the transition operators for both factual and hallucinated regimes and define a differential residual score based on their respective prediction errors. This approach enables low-cost hallucination detection in a single-sample pass, avoiding the need for secondary sampling or external grounding. Extensive testing across three data benchmarks demonstrates that our method achieves state-of-the-art performance with reduced resource overhead.
Machine Learning (cs.LG); Dynamical Systems (math.DS)
arXiv:2605.05134 [cs.LG]
(or arXiv:2605.05134v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2605.05134 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 230 cuvinte
Subiecte conexe
Presentation: Beyond Prompting: Context Engineering and Memory Management for AI Systems at Scale
InfoQ · 10 iun. 2026
[AINews] not much happened today
Latent Space · 18 sept. 2026
ServiceNow AI Control Tower: Selling a penthouse while building the frame
SiliconANGLE · 17 oct. 2026
Problems of Proliferation
LessWrong · 29 sept. 2026