PAIR: Against Inference-Time Attacks on Fine-Tuned LLMs via Path-Level Inference Randomization

Summary

Researchers propose Path-level Inference Randomization (PAIR), a defense for fine-tuned LLMs that randomizes computation inside LoRA adapters at inference time instead of filtering inputs and outputs. PAIR inserts generated shadow weights into feature-sensitive layers to create a stochastic inference network. Across evaluated jailbreak, prompt-stealing and backdoor attacks, the authors report reduced attack metrics with small changes in held-out task performance.

Related items

• MediumCVE-2026-118760: Kubeflow Pipelines artifact routes allow cross-namespace artifact readsSimilar attack · NVD/CVE Database
• MediumCVE-2026-104314: IBM Langflow OSS exposes stored credentials to flow readers via APISimilar attack · NVD/CVE Database
• InfoA Survey of What to Share in Federated Learning: Perspectives on Model Utility, Privacy Leakage, and Communication EfficiencySimilar attack · ACM Digital Library (TOPS, DTRAP, CSUR)
• HighCVE-2026-115415 - Secret masking bypass in the environment tool in Strands Agents ToolsSimilar attack · AWS Security Bulletins
• MediumAttackers built a ‘Skynet’ AI agent that hijacks exposed models to run cyberattacksSimilar attack · CSO Online