Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 8 of 8 for “"Prompt Injection Attacks"”.
-
GuardRAG: Adversarial Preference Training Against Indirect Prompt Injection Attacks
… risks through the retrieval pipeline. Indirect Prompt Injections (IPIs) occur when hidden instructions inside retrieved documents enter a model’s context as trusted evidence, exploiting weak guardrails and leading to unauthorized data leaks, policy overrides, or attacker-controlled behavior. …
-
Optimized Multi-Agent Defense Pipeline Against Advanced Prompt Injection Attacks
… in production remain critically vulnerable to prompt-injection attacks. These attacks embed adversarial inputs that hijack the model into leaking data, taking unauthorized actions, or producing harmful content. Existing multi-agent defenses handle static, known attacks well but leave three …
-
Instruction-Hierarchy Violations and Robustness of Large Language Models to Prompt Injection Attacks
… of modern Large Language Models (LLMs) against prompt injection attacks. Prompt injection is formalized as a violation of instruc tion hierarchy, wherein untrusted user-provided input overrides, modi es, or circumvents trusted system-level instructions. Rather than relying on informal …
-
A MIDDLEWARE SYSTEM FOR DETECTING AND BLOCKING MALICIOUS PROMPT INJECTION ATTACKS IN LLM-BASED APPLICATIONS
Prompt injections can redirect a large language model (LLM) from an application’s intended instructions. In this project, a local middleware system was developed that checks user prompts before passing them on to the model that generates responses. Three detection methods were compared: a …
-
InjectBench: An Indirect Prompt Injection Benchmarking Framework
… it also introduces the risk of indirect prompt injection attacks. In such scenarios, an attacker embeds malicious instructions within the retrieved third party data, which when processed by the LLM, can generate harmful and untruthful outputs for an unsuspecting user. Although previous …
-
Towards Private and Trustworthy Generative Models
… open-generation settings. Next, we present Hard Prompts Made Easy (PEZ), a gradient-based discrete optimization framework that automates the discovery of adversarial prompts, revealing vulnerabilities in safety-aligned models. This work provides a systematic framework for auditing content filters …
-
Lamb: An advanced architecture for secure agents - Extending agentic AI utility while maintaining strong security guarantees through information flow control
Indirect prompt injection attacks are a known and severe problem for the emerging technology of agentic AI. To mitigate them, existing deterministic agents rely on the Dual LLM pattern that disallows untrusted sources from influencing control flow. The pattern prevents execution of benign subtasks …
-
Security and Privacy of Machine Learning Systems
… the attack surface, raising the cost of some attacks while widening the scope of others, but it can also be where a defense is built. In Part I, we study predictive systems and show that a deployed classifier is never the bare function studied in most prior work. We first show that, when a …