AI RESEARCH
The Power of Order: Fooling LLMs with Adversarial Table Permutations
arXiv CS.LG
•
ArXi:2605.00445v1 Announce Type: new Large Language Models have achieved remarkable success and are increasingly deployed in critical applications involving tabular data, such as Table Question Answering. However, their robustness to the structure of this input remains a critical, unaddressed question. This paper nstrates that modern LLMs exhibit a significant vulnerability to the layout of tabular data.