model-hijacking

Live1

1. live
   Hijacking a Trusted LLM in Quantized Deployments via Disguised Adapters

   Research · since 13 Nov 2026

   Researchers demonstrated that third-party LoRA adapters can pass audits on unquantized base models while successfully hijacking decisions after quantization. The attack requires no modifications to the trusted base model.

   AI

Earlier0

Nothing closed yet.