Non Verbis, Sed Rebus: Large Language Models are Weak Solvers of Italian Rebuses

Abstract

Rebuses are puzzles requiring constrained multi-step reasoning to identify a hidden phrase from a set of images and letters. In this work, we introduce a large collection of verbalized rebuses for the Italian language and use it to assess the rebus-solving capabilities of state-of-the-art large language models. While general-purpose systems such as LLaMA-3 and GPT-4o perform poorly on this task, ad-hoc fine-tuning seems to improve models' performance. However, we find that performance gains from training are largely motivated by memorization. Our results suggest that rebus solving remains a challenging test bed to evaluate large language models' linguistic proficiency and sequential instruction-following skills.

Code: https://github.com/gsarti/verbalized-rebus. Artifacts: https://huggingface.co/collections/gsarti/verbalized-rebus-clic-it-2024-66ab8f11cb04e68bdf4fb028