LoREnc: Low-Rank Encryption for Securing Foundation Models and LoRA Adapters

Hugging Face Daily Papers Papers

Summary

LoREnc is a training-free framework that secures foundation models and LoRA adapters via spectral truncation and compensation, preventing unauthorized model recovery while maintaining performance for authorized users. Accepted at ICIP 2026.

Foundation models and low-rank adapters enable efficient on-device generative AI but raise risks such as intellectual property leakage and model recovery attacks. Existing defenses are often impractical because they require retraining or access to the original dataset. We propose LoREnc, a training-free framework that secures both FMs and adapters via spectral truncation and compensation. LoREnc suppresses dominant low-rank components of FM weights, compensates for the missing information in authorized adapters, and further applies orthogonal reparameterization to obscure structural fingerprints of the protected adapter. Unauthorized users produce structurally collapsed outputs, while authorized users recover exact performance. Experiments demonstrate that LoREnc provides strong protection against model recovery with under 1% computational overhead.
Original Article
View Cached Full Text

Cached at: 05/23/26, 02:23 AM

Paper page - LoREnc: Low-Rank Encryption for Securing Foundation Models and LoRA Adapters

Source: https://huggingface.co/papers/2605.13163

Abstract

LoREnc secures foundation models and low-rank adapters through spectral truncation and compensation techniques that prevent unauthorized model recovery while maintaining performance for authorized users.

Foundation modelsandlow-rank adaptersenable efficient on-device generative AI but raise risks such as intellectual property leakage andmodel recovery attacks. Existing defenses are often impractical because they require retraining or access to the original dataset. We propose LoREnc, a training-free framework that secures both FMs and adapters viaspectral truncationandcompensation. LoREnc suppresses dominant low-rank components of FM weights, compensates for the missing information in authorized adapters, and further appliesorthogonal reparameterizationto obscure structural fingerprints of the protected adapter. Unauthorized users produce structurally collapsed outputs, while authorized users recover exact performance. Experiments demonstrate that LoREnc provides strong protection against model recovery with under 1%computational overhead.

View arXiv pageView PDFAdd to collection

Community

Paper author

Paper submitter

about 1 hour ago

edited about 1 hour ago LoREnc protects foundation model weights from extraction by injecting noise into the FM that cancels out in the LoRA adapter — the fine-tuned output stays intact while FM weights remain obfuscated. Accepted at ICIP 2026.

Upload images, audio, and videos by dragging in the text input, pasting, orclicking here.

Tap or paste here to upload images

Get this paper in your agent:

hf papers read 2605\.13163

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2605.13163 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2605.13163 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2605.13163 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Hybrid-LoRA: Bridging Full Fine-Tuning and Low-Rank Adaptation for Post-Training

arXiv cs.LG

Hybrid-LoRA proposes a framework that selectively applies full fine-tuning to a small subset of modules while using LoRA for the rest, achieving performance near full fine-tuning with significantly lower computational cost. Experiments show improvements of up to 5.65% over existing parameter-efficient baselines.

JumpLoRA: Sparse Adapters for Continual Learning in Large Language Models

arXiv cs.CL

JumpLoRA introduces a novel sparse adapter framework for continual learning in LLMs using JumpReLU gating to dynamically isolate task parameters and prevent catastrophic forgetting. The method enhances LoRA-based approaches and outperforms state-of-the-art continual learning methods like ELLA.