analiza aiqbench
This research introduces UniGuardian, a novel, training-free detection framework that unifies the identification of prompt injection, backdoor, and adversarial attacks under the concept of Prompt Trigger Attacks. By leveraging shared mechanisms to measure output distribution shifts from structured perturbations, the system enables simultaneous attack detection and text generation within a single forward pass, offering an efficient defense strategy for LLM security.
UniGuardian: A Unified Defense for Detecting Prompt Injection, Backdoor Attacks and Adversarial Attacks in Large Language Models
arXiv:2502.13141v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) are vulnerable to attacks like prompt injection, backdoor attacks, and adversarial attacks, which manipulate prompts or models to generate harmful outputs. In this paper, departing from traditional deep learning attack paradigms, we explore their intrinsic relationship and collectively term them Prompt Trigger Attacks (PTA). This raises a key question: Given a prompt, can we tell whether a hidden trigger is steering the model's behavior? We propose UniGuardian, to the best of our knowledge the first training-free LLM detector to jointly detect successfully activated prompt injection, backdoor, and adversarial attacks without knowing the attack type. Its shared mechanism measures how structured prompt perturbations shift the model's output distribution. Additionally, we introduce a single-forward strategy to optimize the detection pipeline, enabling simultaneous attack detection and text generation within a shared batched forward pass at each decoding step. Our experiments confirm that UniGuardian accurately and efficiently identifies trigger-activated prompts in LLMs.
Abstract:Large Language Models (LLMs) are vulnerable to attacks like prompt injection, backdoor attacks, and adversarial attacks, which manipulate prompts or models to generate harmful outputs. In this paper, departing from traditional deep learning attack paradigms, we explore their intrinsic relationship and collectively term them Prompt Trigger Attacks (PTA). This raises a key question: Given a prompt, can we tell whether a hidden trigger is steering the model's behavior? We propose UniGuardian, to the best of our knowledge the first training-free LLM detector to jointly detect successfully activated prompt injection, backdoor, and adversarial attacks without knowing the attack type. Its shared mechanism measures how structured prompt perturbations shift the model's output distribution. Additionally, we introduce a single-forward strategy to optimize the detection pipeline, enabling simultaneous attack detection and text generation within a shared batched forward pass at each decoding step. Our experiments confirm that UniGuardian accurately and efficiently identifies trigger-activated prompts in LLMs.
25 Pages, 13 Figures, 11 Tables. Accepted to Findings of AACL-IJCNLP 2026. Keywords: Attack Defending, Security, Prompt Injection, Backdoor Attacks, Adversarial Attacks, Prompt Trigger Attacks
Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2502.13141 [cs.CL]
(or arXiv:2502.13141v2 [cs.CL] for this version)
Text integral sub licența deschisă a publicației · 195 cuvinte
Subiecte conexe
Universal Defenses for Tool-Integrated LLM Agents Against Adversarial Attacks
arXiv · 16 sept. 2026
Cordyceps: Covert Control Attacks on LLMs via Data Poisoning
arXiv · 16 iun. 2026
Presentation: Exploding GPUs
InfoQ · 16 oct. 2026
Defusing Explosive Prompts: Understanding and Preventing Trigger-Based Prompt Injections in LLM Agents
arXiv · 22 sept. 2026