Year: 2024
Type: article
Abstract: The recent surge in popularity of LLMs has led to increased interest in them and to extensive research and evaluation of their reasoning abilities. Induction, deduction, and abduction represent different but interrelated aspects of logical thinking. While each approach serves unique purposes, together they contribute significantly to understanding complex systems and solving complex problems. These reasoning approaches are fundamental in artificial intelligence. Each of them serves different purposes and exhibits specific strengths. In this study, an attempt was made to compare the inductive, deductive and abductive reasoning of three popular LLMs (GPT-4, PaLM-2, LLaMa-2).
Authors: Desislava Baeva, Galina Ivanova
Institution: Angel Kanchev University of Ruse
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
Funder: European Commission
Award: G1562245063
OpenAlex ID: W4405790665