Working Memory Capacity of ChatGPT: An Empirical Study
Dongyu Gong, Xingchen Wan, Dingmin Wang
Abstract
Working memory is a critical aspect of both human intelligence and artificial intelligence, serving as a workspace for the temporary storage and manipulation of information. In this paper, we systematically assess the working memory capacity of ChatGPT, a large language model developed by OpenAI, by examining its performance in verbal and spatial n-back tasks under various conditions. Our experiments reveal that ChatGPT has a working memory capacity limit strikingly similar to that of humans. Furthermore, we investigate the impact of different instruction strategies on ChatGPT's performance and observe that the fundamental patterns of a capacity limit persist. From our empirical findings, we propose that n-back tasks may serve as tools for benchmarking the working memory capacity of large language models and hold potential for informing future efforts aimed at enhancing AI working memory.
BibTeX
@article{Gong_Wan_Wang_2024, title={Working Memory Capacity of ChatGPT: An Empirical Study}, volume={38}, url={https://ojs.aaai.org/index.php/AAAI/article/view/28868}, DOI={10.1609/aaai.v38i9.28868}, abstractNote={Working memory is a critical aspect of both human intelligence and artificial intelligence, serving as a workspace for the temporary storage and manipulation of information. In this paper, we systematically assess the working memory capacity of ChatGPT, a large language model developed by OpenAI, by examining its performance in verbal and spatial n-back tasks under various conditions. Our experiments reveal that ChatGPT has a working memory capacity limit strikingly similar to that of humans. Furthermore, we investigate the impact of different instruction strategies on ChatGPT’s performance and observe that the fundamental patterns of a capacity limit persist. From our empirical findings, we propose that n-back tasks may serve as tools for benchmarking the working memory capacity of large language models and hold potential for informing future efforts aimed at enhancing AI working memory.}, number={9}, journal={Proceedings of the AAAI Conference on Artificial Intelligence}, author={Gong, Dongyu and Wan, Xingchen and Wang, Dingmin}, year={2024}, month={Mar.}, pages={10048-10056} }