SemBench: A Universal Semantic Framework for LLM Evaluation

TL;DR

SemBench is a scalable, language-independent framework that automatically generates synthetic benchmarks to evaluate the semantic understanding of large language models across multiple languages.

Contribution

Introduces SemBench, a novel, resource-efficient method for assessing LLM semantic competence using dictionary definitions and sentence encoders, enabling cross-lingual evaluation.

Findings

SemBench rankings correlate well with WiC datasets.

Few examples suffice for stable model ranking.

Framework is scalable and language-independent.

Abstract

Recent progress in Natural Language Processing (NLP) has been driven by the emergence of Large Language Models (LLMs), which exhibit remarkable generative and reasoning capabilities. However, despite their success, evaluating the true semantic understanding of these models remains a persistent challenge. Traditional benchmarks such as Word-in-Context (WiC) effectively probe this capability, but their creation is resource-intensive and often limited to high-resource languages. In this paper, we introduce SemBench, a framework for automatically generating synthetic benchmarks that assess the semantic competence of LLMs using only dictionary sense definitions and a sentence encoder. This approach eliminates the need for curated example sentences, making it both scalable and language-independent. We evaluate SemBench in three languages (English, Spanish, and Basque) spanning different levels of linguistic resources, and across a wide range of LLMs. Our results show that rankings derived from SemBench strongly correlate with those obtained from standard WiC datasets. Furthermore, our analysis demonstrates that only a small number of examples is required to achieve stable and meaningful rankings. Overall, SemBench provides a lightweight, adaptable, and data-efficient framework for cross-lingual evaluation of semantic understanding in LLMs.