Hijacking a Trusted LLM in Quantized Deployments via Disguised Adapters

# Hijacking a Trusted LLM in Quantized Deployments via Disguised Adapters

https://gcdradar.vercel.app/incidents/hijacking-a-trusted-llm-in-quantized-deployments-via-disguis · AIR-2026-1960 · live

- Lane: Research
- Severity: 1 (Research / proof of concept): Demonstrated research attack with no known real-world exploitation.
- Fix status: Fix status unknown
- First reported: 2026-11-13
- AI role: Attack on AI

## Summary

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.

## What to do

No action recorded; read the sources below and the vendor's advisory.

## Latest updates

- 2026-11-13: Hijacking a Trusted LLM in Real-World Quantized Deployments Only Requires a Disguised Adapter (https://arxiv.org/abs/2611.14978)

## Sources

- 2026-11-13 arXiv: Hijacking a Trusted LLM in Real-World Quantized Deployments Only Requires a Disguised Adapter https://arxiv.org/abs/2611.14978

Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.