Pun Unintended: LLMs and the Illusion of Humor Understanding

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zangari, Alessandro, Marcuzzo, Matteo, Albarelli, Andrea, Pilehvar, Mohammad Taher, Camacho-Collados, Jose
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918144731250688
author Zangari, Alessandro
Marcuzzo, Matteo
Albarelli, Andrea
Pilehvar, Mohammad Taher
Camacho-Collados, Jose
author_facet Zangari, Alessandro
Marcuzzo, Matteo
Albarelli, Andrea
Pilehvar, Mohammad Taher
Camacho-Collados, Jose
contents Puns are a form of humorous wordplay that exploits polysemy and phonetic similarity. While LLMs have shown promise in detecting puns, we show in this paper that their understanding often remains shallow, lacking the nuanced grasp typical of human interpretation. By systematically analyzing and reformulating existing pun benchmarks, we demonstrate how subtle changes in puns are sufficient to mislead LLMs. Our contributions include comprehensive and nuanced pun detection benchmarks, human evaluation across recent LLMs, and an analysis of the robustness challenges these models face in processing puns.
format Preprint
id arxiv_https___arxiv_org_abs_2509_12158
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Pun Unintended: LLMs and the Illusion of Humor Understanding
Zangari, Alessandro
Marcuzzo, Matteo
Albarelli, Andrea
Pilehvar, Mohammad Taher
Camacho-Collados, Jose
Computation and Language
Artificial Intelligence
68T50
I.2.7
Puns are a form of humorous wordplay that exploits polysemy and phonetic similarity. While LLMs have shown promise in detecting puns, we show in this paper that their understanding often remains shallow, lacking the nuanced grasp typical of human interpretation. By systematically analyzing and reformulating existing pun benchmarks, we demonstrate how subtle changes in puns are sufficient to mislead LLMs. Our contributions include comprehensive and nuanced pun detection benchmarks, human evaluation across recent LLMs, and an analysis of the robustness challenges these models face in processing puns.
title Pun Unintended: LLMs and the Illusion of Humor Understanding
topic Computation and Language
Artificial Intelligence
68T50
I.2.7
url https://arxiv.org/abs/2509.12158