Abubakar Zakari

Abstract Large Language Models (LLMs) are increasingly being applied to time-series forecasting, giving rise to a class of models referred to as time-series LLMs. While these models achieve competitive predictive accuracy, their reliability and structural consistency remain insufficiently understood. These models may produce forecasts that are numerically accurate on average yet statistically or temporally implausible, a phenomenon referred to as hallucination.

1. Introduction The miniaturization of sensing, computing, networking, and communication technologies in urban spaces has led to the development of new technological ecosystems for smart cities []. In present times, city governments are enabling pervasive smart services across all application domains [], including healthcare, transportation, clean and green technologies, entertainment and leisure facilities, and crowd management, among others [].