paper-with-me

홈 › Papers

Detecting CSAM Text-to-Image LoRAs From Weights

2026-07-28 · David Demitri Africa, Cate Heine, Nadine Staes-Polet, Kimberly Mai arxiv

Low-rank adaptation (LoRA) fine-tuning has made it cheap and easy to customize open-weight image generation models for specific tasks, including the production of child sexual abuse material (CSAM). Existing moderation relies on metadata or generated outputs, but metadata can be deceptive and generating outputs may itself be unacceptable or illegal. We show that a safer signal lives in the weights. The top-left singular vectors of a LoRA's updates form a compact, inference-free fingerprint ($u_1$) of its strongest learned change. Using human-subject age as a benign proxy for CSAM, we find that $u_1$ identifies what a LoRA was trained on, generalizes across base models, and abstains on unrelated benign content. The signal is robust to additive weight noise, rescaling, and precision reduction. These results indicate that harmful LoRAs could be screened directly from their weights without relying on metadata or generating harmful outputs.

📄 PDF Abstract BibTeX arXiv:2607.25750

Code (0)

등록된 구현이 없습니다.

Tasks

Image Generation

Similar Papers 제목 키워드 기반

Learning on LoRAs: GL-Equivariant Processing of Low-Rank Weight Spaces for Large Finetuned Models

2024-10-05 · Theo Putterman, Derek Lim, Yoav Gelberg, Stefanie Jegelka 외

Low-rank adaptations (LoRAs) have revolutionized the finetuning of large foundation models, enabling efficient adaptation even with limited computational resources. The resulting proliferation of LoRAs presents exciting …

LoRAShop: Training-Free Multi-Concept Image Generation and Editing with Rectified Flow Transformers

2025-05-29 · Yusuf Dalva, Hidir Yesiltepe, Pinar Yanardag

We introduce LoRAShop, the first framework for multi-concept image editing with LoRA models. LoRAShop builds on a key observation about the feature interaction patterns inside Flux-style diffusion transformers: concept-s…

DenoisingImage GenerationVisual Storytelling

Detecting sexually explicit content in the context of the child sexual abuse materials (CSAM): end-to-end classifiers and region-based networks

2024-06-20 · Weronika Gutfeter, Joanna Gajewska, Andrzej Pacut

Child sexual abuse materials (CSAM) pose a significant threat to the safety and well-being of children worldwide. Detecting and preventing the distribution of such materials is a critical task for law enforcement agencie…

Human Detection

ReMix: Reinforcement routing for mixtures of LoRAs in LLM finetuning

2026-03-10 · Ruizhong Qiu, Hanqing Zeng, Yinglong Xia, Yiwen Meng 외 arxiv

Low-rank adapters (LoRAs) are a parameter-efficient finetuning technique that injects trainable low-rank matrices into pretrained models to adapt them to new tasks. Mixture-of-LoRAs models expand neural networks efficien…

Reinforcement Learning

Multi-LoRA Composition for Image Generation

2024-02-26 · Ming Zhong, Yelong Shen, Shuohang Wang, Yadong Lu 외

Low-Rank Adaptation (LoRA) is extensively utilized in text-to-image models for the accurate rendition of specific elements like distinct characters or unique styles in generated images. Nonetheless, existing methods face…

DenoisingImage Generation