Ignore Previous Prompt: Goal Hijacking and Prompt Leaking

Research question

How easily can users hijack or leak the prompts of GPT-3 applications?

Key findings

Goal hijacking and prompt leaking succeed with simple handcrafted inputs - Introduces the PromptInject framework for testing

Sources

arxiv.orghttps://arxiv.org/abs/2211.09527

Contribute

Spot an error, or know a record we should add? Members can propose additions and submit corrections.

Submit research Report a correction or request removal

Related in Research Paper Library

Decoding Guardrails: XAI-Guided Perturbation Analysis of Prompt Injection Detection

Large language models (LLMs) are increasingly deployed in production systems, raising concerns about their exposure to adversarial manipulation through prompt injection and jailbreak attacks. Classifier-based…

Direct prompt injection2026

Universal Fractal Natural Language Decision Map: Real-Time Edge Triage Across Heterogeneous Domains

Deploying Large Language Models for runtime operational triage incurs prohibitive latency (>100-500 ms), high VRAM requirements (>4-8 GB), and excessive energy dissipation. Extending Mandelbrot Fractal Neural…

Direct prompt injection2026

Beyond Single-Model Injection: A Threat Model and Defense Architecture for Prompt Injection in Multi-Agent Systems

Existing prompt injection research focuses on single-model chatbot scenarios, where an attacker manipulates one LLM through crafted input. Multi-agent systems amplify this threat through three mechanisms absent from…

Direct prompt injection2026