Evaluation of Prompt Injection Defenses in Large Language Models
For developers deploying LLM-powered applications, this work provides empirical evidence that security boundaries must be enforced in application code rather than relying on the model itself.
The paper evaluates nine defense configurations against adaptive prompt injection attacks on LLMs, finding that only output filtering in application code prevented leaks (zero leaks across 15,000 attacks), while all model-based defenses eventually broke.
LLM-powered applications routinely embed secrets in system prompts, yet models can be tricked into revealing them. We built an adaptive attacker that evolves its strategies over hundreds of rounds and tested it against nine defense configurations across more than 20,000 attacks. Every defense that relied on the model to protect itself eventually broke. The only defense that held was output filtering, which checks the model's responses via hardcoded rules in separate application code before they reach the user, achieving zero leaks across 15,000 attacks. These results demonstrate that security boundaries must be enforced in application code, not by the model being attacked. Until such defenses are verified by tools like Swept AI, AI systems handling sensitive operations should be restricted to internal, trusted personnel.