# Weight Space Representation Learning via Neural Field Adaptation Source: [Weight Space Representation Learning via Neural Field Adaptation]() Collection index: [llms.txt]() Data updated: 2026-06-26T05:00:04.994Z Authors: Zhuoqian Yang, Mathieu Salzmann, Sabine S\\"usstrunk Published: 2025-12-01T15:05:01.000Z ## Abstract We investigate the potential of weights to serve as effective representations, focusing on neural fields. Our key insight is that constraining the optimization space through a pre-trained base model and low-rank adaptation (LoRA) can induce structure in weight space. Across reconstruction, generation, and analysis tasks on 2D and 3D data, we find that multiplicative LoRA weights achieve high representation quality while exhibiting distinctiveness and semantic structure. When used with latent diffusion models, multiplicative LoRA weights enable higher-quality generation than existing weight-space methods. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Weight Space Generation on FFHQ 2D](): Weight Space Generation; FFHQ 2D; FD: 0.073. - [Weight space generation on ShapeNet Airplane](): Weight Space Generation; ShapeNet airplane; mMD: 1.89. - [Weight Space Generation on ShapeNet Multi](): Weight Space Generation; ShapeNet Multi; mMD: 5.52. ## Original sources and implementation [Original paper]() [PDF]()