Large Language Models (LLMs) effectiveness is usually evaluated by means of benchmarks such as MMLU, ARC-C, or HellaSwag, where questions are presented in their original wording, thus in a fixed, standardized format. However, real-world applications involve linguistic variability, requiring models to maintain their effectiveness across diverse rewordings of the same question or query. In this study, we systematically assess the robustness of LLMs to paraphrased benchmark questions and investigate whether benchmark-based evaluations provide a reliable measure of model capabilities. We systematically generate various paraphrases of all the questions across six different common benchmarks, and measure the resulting variations in effectiveness of 34 state-of-the-art LLMs, of different size and effectiveness. Our findings reveal that while LLM rankings remain relatively stable across paraphrased inputs, absolute effectiveness scores change, and decline significantly. This suggests that LLMs struggle with linguistic variability, raising concerns about their generalization abilities and evaluation methodologies. Furthermore, the observed performance drop challenges the reliability of benchmark-based evaluations, indicating that high benchmark scores may not fully capture a model’s robustness to real-world input variations. We discuss the implications of these findings for LLM evaluation methodologies, emphasizing the need for robustness-aware benchmarks that better reflect practical deployment scenarios.

Patient Information Leaflets (PILs) provide essential information about medication usage, side effects, precautions, and interactions, making them a valuable resource for Question Answering (QA) systems in healthcare. However, no dedicated benchmark currently exists to evaluate QA systems specifically on PILs, limiting progress in this domain. To address this gap, we introduce a fact-supported synthetic benchmark composed of multiple-choice questions and answers generated from real PILs. We construct the benchmark using a fully automated pipeline that leverages multiple Large Language Models (LLMs) to generate diverse, realistic, and contextually relevant question-answer pairs. The benchmark is publicly released as a standardized evaluation framework for assessing the ability of LLMs to process and reason over PIL content. To validate its effectiveness, we conduct an initial evaluation with state-of-the-art LLMs, showing that the benchmark presents a realistic and challenging task, making it a valuable resource for advancing QA research in the healthcare domain.

This study challenges the prevailing approach of measuring political leanings in Large Language Models (LLMs) through direct questioning. By extensively testing LLMs with original, positively and negatively paraphrased Political Compass questions we demonstrate that LLMs do not consistently reveal their political biases in response to standard questions. Our findings indicate that LLMs' political orientations are elusive, easily influenced by subtle changes in phrasing and context. This study underscores the limitations of direct questioning in accurately measuring the political biases of LLMs and emphasizes the necessity for more refined and effective approaches to understand their true political stances.