← 研踪 PaperTrailarXiv:2611.03752 · v1
LoRA-Patch: Defending Against Backdoored LoRA Adapters via Directional Steering
首次提交 2026-11-02 16:44
版本更新 2026-11-02 16:44
雷达观测 2026-11-04 23:42
摘要
Third-party LoRA adapters introduce serious backdoor risks: malicious adapters can maintain normal task performance while embedding hidden, trigger-activated behaviors, making them difficult to detect through standard utility evaluations. Existing removal-based defenses typically sanitize backdoored models by updating, pruning, distilling, or mixing model parameters. However, these approaches are misaligned with adapter-based deployment, where the trusted base model is expected to remain frozen. In this work, we propose LoRA-Patch, a lightweight backdoor defense tailored to LoRA adapters. Our key insight is that backdoor injection distorts the directional geometry of a poisoned adapter relative to its trusted base model. Building on this insight, LoRA-Patch suppresses trigger-activated behaviors by realigning adapter directions through layer-wise steering, without modifying the base model weights. Extensive experiments across five backdoor attacks and three models show that LoRA-Patch effectively mitigates trigger-activated backdoors, outperforming existing baselines while preserving benign task performance. Moreover, LoRA-Patch is computationally efficient, constructing patches within 20 seconds across all evaluated models and adapters. Our code is publicly available at https://github.com/FOXVEDA/LoRA-Patch.
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-02 16:44 · 最新观测版本
LoRA-Patch: Defending Against Backdoored LoRA Adapters via Directional Steering
Third-party LoRA adapters introduce serious backdoor risks: malicious adapters can maintain normal task performance while embedding hidden, trigger-activated behaviors, making them difficult to detect through standard utility evaluations. Existing removal-based defenses typically sanitize backdoored models by updating, pruning, distilling, or mixing model parameters. However, these approaches are misaligned with adapter-based deployment, where the trusted base model is expected to remain frozen. In this work, we propose LoRA-Patch, a lightweight backdoor defense tailored to LoRA adapters. Our key insight is that backdoor injection distorts the directional geometry of a poisoned adapter relative to its trusted base model. Building on this insight, LoRA-Patch suppresses trigger-activated behaviors by realigning adapter directions through layer-wise steering, without modifying the base model weights. Extensive experiments across five backdoor attacks and three models show that LoRA-Patch effectively mitigates trigger-activated backdoors, outperforming existing baselines while preserving benign task performance. Moreover, LoRA-Patch is computationally efficient, constructing patches within 20 seconds across all evaluated models and adapters. Our code is publicly available at https://github.com/FOXVEDA/LoRA-Patch.
查看此版本解读 →查看此版本原文 ↗