Do Large Language Models Solve ARC Visual Analogies Like People Do?

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Opiełka, Gustaw, Rosenbusch, Hannes, Vijverberg, Veerle, Stevenson, Claire E.
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911873988820992
author Opiełka, Gustaw
Rosenbusch, Hannes
Vijverberg, Veerle
Stevenson, Claire E.
author_facet Opiełka, Gustaw
Rosenbusch, Hannes
Vijverberg, Veerle
Stevenson, Claire E.
contents The Abstraction Reasoning Corpus (ARC) is a visual analogical reasoning test designed for humans and machines (Chollet, 2019). We compared human and large language model (LLM) performance on a new child-friendly set of ARC items. Results show that both children and adults outperform most LLMs on these tasks. Error analysis revealed a similar "fallback" solution strategy in LLMs and young children, where part of the analogy is simply copied. In addition, we found two other error types, one based on seemingly grasping key concepts (e.g., Inside-Outside) and the other based on simple combinations of analogy input matrices. On the whole, "concept" errors were more common in humans, and "matrix" errors were more common in LLMs. This study sheds new light on LLM reasoning ability and the extent to which we can use error analyses and comparisons with human development to understand how LLMs solve visual analogies.
format Preprint
id arxiv_https___arxiv_org_abs_2403_09734
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Do Large Language Models Solve ARC Visual Analogies Like People Do?
Opiełka, Gustaw
Rosenbusch, Hannes
Vijverberg, Veerle
Stevenson, Claire E.
Computation and Language
Artificial Intelligence
The Abstraction Reasoning Corpus (ARC) is a visual analogical reasoning test designed for humans and machines (Chollet, 2019). We compared human and large language model (LLM) performance on a new child-friendly set of ARC items. Results show that both children and adults outperform most LLMs on these tasks. Error analysis revealed a similar "fallback" solution strategy in LLMs and young children, where part of the analogy is simply copied. In addition, we found two other error types, one based on seemingly grasping key concepts (e.g., Inside-Outside) and the other based on simple combinations of analogy input matrices. On the whole, "concept" errors were more common in humans, and "matrix" errors were more common in LLMs. This study sheds new light on LLM reasoning ability and the extent to which we can use error analyses and comparisons with human development to understand how LLMs solve visual analogies.
title Do Large Language Models Solve ARC Visual Analogies Like People Do?
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2403.09734