Enregistré dans:
Détails bibliographiques
Auteurs principaux: Galanti, Liane, Baron, Ethan
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2407.18968
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866909270180626432
author Galanti, Liane
Baron, Ethan
author_facet Galanti, Liane
Baron, Ethan
contents In this project, we test the effectiveness of Large Language Models (LLMs) on the Abstraction and Reasoning Corpus (ARC) dataset. This dataset serves as a representative benchmark for testing abstract reasoning abilities, requiring a fundamental understanding of key concepts such as object identification, basic counting, and elementary geometric principles. Tasks from this dataset are converted into a prompt-based format for evaluation. Initially, we assess the models' potential through a Zero-shot approach. Subsequently, we investigate the application of the Chain-of-Thought (CoT) technique, aiming to determine its role in improving model performance. Our results suggest that, despite the high expectations placed on contemporary LLMs, these models still struggle in non-linguistic domains, even when dealing with simpler subsets of the ARC dataset. Our study is the first to concentrate on the capabilities of open-source models in this context. The code, dataset, and prompts supporting this project's findings can be found in our GitHub repository, accessible at: https://github.com/Lianga2000/LLMsOnARC.
format Preprint
id arxiv_https___arxiv_org_abs_2407_18968
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Intelligence Analysis of Language Models
Galanti, Liane
Baron, Ethan
Artificial Intelligence
In this project, we test the effectiveness of Large Language Models (LLMs) on the Abstraction and Reasoning Corpus (ARC) dataset. This dataset serves as a representative benchmark for testing abstract reasoning abilities, requiring a fundamental understanding of key concepts such as object identification, basic counting, and elementary geometric principles. Tasks from this dataset are converted into a prompt-based format for evaluation. Initially, we assess the models' potential through a Zero-shot approach. Subsequently, we investigate the application of the Chain-of-Thought (CoT) technique, aiming to determine its role in improving model performance. Our results suggest that, despite the high expectations placed on contemporary LLMs, these models still struggle in non-linguistic domains, even when dealing with simpler subsets of the ARC dataset. Our study is the first to concentrate on the capabilities of open-source models in this context. The code, dataset, and prompts supporting this project's findings can be found in our GitHub repository, accessible at: https://github.com/Lianga2000/LLMsOnARC.
title Intelligence Analysis of Language Models
topic Artificial Intelligence
url https://arxiv.org/abs/2407.18968