Author name: Hananel Hazan, Yanbo Zhang, Benedikt Hartl, Michael Levin

arXiv:2604.08749v2 Announce Type: replace
Abstract: How many of a neural network’s parameters actually encode task-specific information? We investigate this question with LottaLoRA, a training paradigm in which every backbone weight is drawn at random…

arXiv:2604.08749v1 Announce Type: new
Abstract: How many of a neural network’s parameters actually encode task-specific information? We investigate this question with LottaLoRA, a training paradigm in which every backbone weight is drawn at random and…