Explaining Large Language Models Decisions Using Shapley Values
Computer Science > Computation and Language
arXiv:2404.01332 (cs)
[Submitted on 29 Mar 2024 (v1), last revised 12 Nov 2024 (this version, v3)]
View PDF
Abstract:The emergence of large language models (LLMs) has opened up exciting possibilities for simulating human behavior and cognitive processes, with potential applications in various domains, including marketing research and consumer behavior analysis. However, the validity of utilizing LLMs as stand-ins for human subjects remains uncertain du...
Read more at arxiv.org