A Fine-Tuning Data Recovery Attack on Generative Language Models via Backdooring

Summary

Researchers present Lure, a backdoor-based data recovery attack on generative language models. By injecting a backdoor into the source code of an open-source GLM, an adversary can make the model memorize fine-tuning data and regenerate it via crafted hash prompts. Evaluations report a 45%-68% data recovery rate while the attack remains stealthy and evades existing defenses.

Related items

• HighGHSA-6wjp-v33h-5cvq: PraisonAI: AgentOS defaults to network-exposed no-auth mode, allowing unauthenticated agent invocation and instruction disclosureSimilar attack · GitHub Advisory Database
• HighCVE-2026-101998: Docker Sandboxes fail open when masking credentials in proxy responsesSimilar attack · NVD/CVE Database
• InfoStepSecurity Now Inventories AI Agent Skills in Your GitHub Repositories and on Developer Machines Similar attack · StepSecurity Blog
• LowCVE-2026-107288: Pydantic AI web fetch bypasses blocked_domains via hostname variantsSimilar attack · NVD/CVE Database
• LowOAuth grants pile up faster than you can review them. Here's how to keep up.Similar attack · BleepingComputer