Talking the Talk Does Not Entail Walking the Walk: On the Limits of Large Language Models in Lexical Entailment Recognition

TL;DR

This study evaluates eight large language models' ability to recognize lexical entailment among verbs, revealing moderate success with potential improvements from few-shot prompting but highlighting the ongoing challenge of perfect accuracy.

Contribution

The paper systematically assesses LLMs' performance on lexical entailment recognition for verbs using various prompting strategies and datasets, revealing current limitations and avenues for future research.

Findings

Models achieve moderate performance in lexical entailment recognition.

Few-shot prompting can improve model accuracy.

Perfect lexical entailment recognition remains an open challenge.

Abstract

Verbs form the backbone of language, providing the structure and meaning to sentences. Yet, their intricate semantic nuances pose a longstanding challenge. Understanding verb relations through the concept of lexical entailment is crucial for comprehending sentence meanings and grasping verb dynamics. This work investigates the capabilities of eight Large Language Models in recognizing lexical entailment relations among verbs through differently devised prompting strategies and zero-/few-shot settings over verb pairs from two lexical databases, namely WordNet and HyperLex. Our findings unveil that the models can tackle the lexical entailment recognition task with moderately good performance, although at varying degree of effectiveness and under different conditions. Also, utilizing few-shot prompting can enhance the models' performance. However, perfectly solving the task arises as an unmet challenge for all examined LLMs, which raises an emergence for further research developments on this topic.