Structure Is Not Enough: Leveraging Behavior for Neural Network Weight Reconstruction

Published: 05 Mar 2025, Last Modified: 05 Mar 2025ICLR 2025 Workshop Weight Space Learning PosterEveryoneRevisionsBibTeXCC BY 4.0
Track: long paper (up to 8 pages)
Keywords: probing-based loss function, self-supervised learning, autoencoders
TL;DR: We show that training weight-space autoencoders with a behavioral loss component improves neural network weights reconstruction and generation
Submission Number: 35
Loading

OpenReview is a long-term project to advance science through improved peer review with legal nonprofit status. We gratefully acknowledge the support of the OpenReview Sponsors. © 2025 OpenReview