Information Anxiety in Large Language Models
This work addresses the problem of LLMs' reliability in handling frequently occurring entities, which is incremental as it builds on prior studies of retrieval capabilities.
The paper investigates how entity popularity affects LLMs' internal reasoning and retrieval, finding that while popular queries lead to early convergence, they also cause retrieved attributes to become less accurate and dissimilar across lexical variations, a phenomenon termed 'information anxiety'.
Large Language Models (LLMs) have demonstrated strong performance as knowledge repositories, enabling models to understand user queries and generate accurate and context-aware responses. Extensive evaluation setups have corroborated the positive correlation between the retrieval capability of LLMs and the frequency of entities in their pretraining corpus. We take the investigation further by conducting a comprehensive analysis of the internal reasoning and retrieval mechanisms of LLMs. Our work focuses on three critical dimensions - the impact of entity popularity, the models' sensitivity to lexical variations in query formulation, and the progression of hidden state representations across LLM layers. Our preliminary findings reveal that popular questions facilitate early convergence of internal states toward the correct answer. However, as the popularity of a query increases, retrieved attributes across lexical variations become increasingly dissimilar and less accurate. Interestingly, we find that LLMs struggle to disentangle facts, grounded in distinct relations, from their parametric memory when dealing with highly popular subjects. Through a case study, we explore these latent strains within LLMs when processing highly popular queries, a phenomenon we term information anxiety. The emergence of information anxiety in LLMs underscores the adversarial injection in the form of linguistic variations and calls for a more holistic evaluation of frequently occurring entities.