arXiv NLP
techCenter
How Perturbations Propagate: A Multi-Level Analysis of Robustness in Large Language Modelstranslating…
arXiv:2609.03322v1 Announce Type: new
Abstract: Language models encounter typos, corrupted text, altered words, and disrupted token order, yet robustness is usually evaluated only through output behavior. We study how six naturalistic and synthetic input perturbations propagate…
Keywords#Language Models#Perturbations Propagate#Robustness#A Multi-Level#A Multi-Level Analysis
Comments
Sign in to join the discussion.
Loading comments…