# Trojan Horse Prompting: Jailbreaking Conversational Multimodal Models by Forging Assistant Message

arXiv: https://arxiv.org/abs/2507.04673 · HTML: https://mlsanity.com/paper/2507.04673/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Jailbreaking, one of 12 directions in LLM safety and security.
- Builds on 21 of its 30 references that are in the map.
- First posted on arXiv 7 Jul 2025.

## Placed in

- LLM safety and security / [Jailbreaking](https://mlsanity.com/safety/jailbreaking/)

## Builds on (most cited first, 21 of 21)

- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Deep reinforcement learning from human preferences](https://mlsanity.com/paper/1706.03741/) (2017)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [Jailbroken: How Does LLM Safety Training Fail?](https://mlsanity.com/paper/2307.02483/) (2023)
- [Red Teaming Language Models with Language Models](https://mlsanity.com/paper/2202.03286/) (2022)
- [FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Prompts](https://mlsanity.com/paper/2311.05608/) (2023)
- [ArtPrompt: ASCII Art-based Jailbreak Attacks against Aligned LLMs](https://mlsanity.com/paper/2402.11753/) (2024)
- [SneakyPrompt: Jailbreaking Text-to-image Generative Models](https://mlsanity.com/paper/2305.12082/) (2023)
- [Agent Smith: A Single Image Can Jailbreak One Million Multimodal LLM Agents Exponentially Fast](https://mlsanity.com/paper/2402.08567/) (2024)
- [ImgTrojan: Jailbreaking Vision-Language Models with ONE Image](https://mlsanity.com/paper/2403.02910/) (2024)
- [Jailbreaking Multimodal Large Language Models via Shuffle Inconsistency](https://mlsanity.com/paper/2501.04931/) (2025)
- [HTS-Attack: Heuristic Token Search for Jailbreaking Text-to-Image Models](https://mlsanity.com/paper/2408.13896/) (2024)
- [Reason2Attack: Jailbreaking Text-to-Image Models via LLM Reasoning](https://mlsanity.com/paper/2503.17987/) (2025)
- [Gradient-based Jailbreak Images for Multimodal Fusion Models](https://mlsanity.com/paper/2410.03489/) (2024)
- [Multimodal Pragmatic Jailbreak on Text-to-image Models](https://mlsanity.com/paper/2409.19149/) (2024)
- [Chain-of-Jailbreak Attack for Image Generation Models via Editing Step by Step](https://mlsanity.com/paper/2410.03869/) (2024)
- [GenBreak: Red Teaming Text-to-Image Generators Using Large Language Models](https://mlsanity.com/paper/2506.10047/) (2025)
- [Adversarial Attacks on Parts of Speech: An Empirical Study in Text-to-Image Generation](https://mlsanity.com/paper/2409.15381/) (2024)

## Built on by (most cited first, 2 of 2)

- [SaFeR-Steer: Evolving Multi-Turn MLLMs via Synthetic Bootstrapping and Feedback Dynamics](https://mlsanity.com/paper/2604.16358/) (2026)
- [Epistemic Policy Divergence in Multi-Turn LLM Contamination: A Protocol-Gradient Investigation](https://mlsanity.com/paper/2609.35308/) (2026)

## Abstract (arXiv)

The rise of conversational interfaces has greatly enhanced LLM usability by leveraging dialogue history for sophisticated reasoning. However, this reliance introduces an unexplored attack surface. This paper introduces Trojan Horse Prompting, a novel jailbreak technique. Adversaries bypass safety mechanisms by forging the model's own past utterances within the conversational history provided to its API. A malicious payload is injected into a model-attributed message, followed by a benign user prompt to trigger harmful content generation. This vulnerability stems from Asymmetric Safety Alignment: models are extensively trained to refuse harmful user requests but lack comparable skepticism towards their own purported conversational history. This implicit trust in its "past" creates a high-impact vulnerability. Experimental validation on Google's Gemini-2.0-flash-preview-image-generation shows Trojan Horse Prompting achieves a significantly higher Attack Success Rate (ASR) than established user-turn jailbreaking methods. These findings reveal a fundamental flaw in modern conversational AI security, necessitating a paradigm shift from input-level filtering to robust, protocol-level validation of conversational context integrity.