Analysis of the Neglect-Zero Effect in Large Language Models

Jin Tanaka, Daiki Matsuoka, Ryoma Kumon +1

We investigate the extent to which the language processing of LLMs resembles human cognitive processes, focusing on a human cognitive bias called the $\text{neglect-zero effect}$…