# Phantasia: Context-Adaptive Backdoors in Vision Language Models

arXiv: https://arxiv.org/abs/2604.08395 · HTML: https://mlsanity.com/paper/2604.08395/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 25 of its 34 references that are in the map.
- First posted on arXiv 9 Apr 2026.

## Placed in

- Multimodal LLMs / vision-language models / [MLLM jailbreaks and adversarial robustness](https://mlsanity.com/mllm/mllm-jailbreaks-and-adversarial-robustness/)
- LLM safety and security / [Backdoor and data poisoning attacks](https://mlsanity.com/safety/backdoor-and-data-poisoning-attacks/)

## Builds on (most cited first, 25 of 25)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [Flamingo: a Visual Language Model for Few-Shot Learning](https://mlsanity.com/paper/2204.14198/) (2022)
- [BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation](https://mlsanity.com/paper/2201.12086/) (2022)
- [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://mlsanity.com/paper/2305.06500/) (2023)
- [MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models](https://mlsanity.com/paper/2304.10592/) (2023)
- [Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering](https://mlsanity.com/paper/1612.00837/) (2016)
- [OK-VQA: A Visual Question Answering Benchmark Requiring External Knowledge](https://mlsanity.com/paper/1906.00067/) (2019)
- [BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain](https://mlsanity.com/paper/1708.06733/) (2017)
- [OpenFlamingo: An Open-Source Framework for Training Large Autoregressive Vision-Language Models](https://mlsanity.com/paper/2308.01390/) (2023)
- [STRIP: A Defence Against Trojan Attacks on Deep Neural Networks](https://mlsanity.com/paper/1902.06531/) (2019)
- [Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attack](https://mlsanity.com/paper/2404.01833/) (2024)
- [ONION: A Simple and Effective Defense Against Textual Backdoor Attacks](https://mlsanity.com/paper/2011.10369/) (2020)
- [Reconstructive Neuron Pruning for Backdoor Defense](https://mlsanity.com/paper/2305.14876/) (2023)
- [Revisiting Backdoor Attacks against Large Vision-Language Models from Domain Shift](https://mlsanity.com/paper/2406.18844/) (2024)
- [Playing the Fool: Jailbreaking LLMs and Multimodal LLMs with Out-of-Distribution Strategy](https://mlsanity.com/paper/2503.20823/) (2025)
- [Test-Time Backdoor Attacks on Multimodal Large Language Models](https://mlsanity.com/paper/2402.08577/) (2024)
- [TrojVLM: Backdoor Attack Against Vision Language Models](https://mlsanity.com/paper/2409.19232/) (2024)
- [Backdoor Defense via Deconfounded Representation Learning](https://mlsanity.com/paper/2303.06818/) (2023)
- [Improving Automatic VQA Evaluation Using Large Language Models](https://mlsanity.com/paper/2310.02567/) (2023)
- [Physical Backdoor Attack can Jeopardize Driving with Vision-Large-Language Models](https://mlsanity.com/paper/2404.12916/) (2024)
- [Backdooring Vision-Language Models with Out-Of-Distribution Data](https://mlsanity.com/paper/2410.01264/) (2024)
- [Stealthy Backdoor Attack in Self-Supervised Learning Vision Encoders for Large Vision Language Models](https://mlsanity.com/paper/2502.18290/) (2025)
- [Backdoor Attack on Vision Language Models with Stealthy Semantic Manipulation](https://mlsanity.com/paper/2506.07214/) (2025)

## Abstract (arXiv)

Recent advances in Vision-Language Models (VLMs) have greatly enhanced the integration of visual perception and linguistic reasoning, driving rapid progress in multimodal understanding. Despite these achievements, the security of VLMs, particularly their vulnerability to backdoor attacks, remains significantly underexplored. Existing backdoor attacks on VLMs are still in an early stage of development, with most current methods relying on generating poisoned responses that contain fixed, easily identifiable patterns. In this work, we make two key contributions. First, we demonstrate for the first time that the stealthiness of existing VLM backdoor attacks has been substantially overestimated. By adapting defense techniques originally designed for other domains (e.g., vision-only and text-only models), we show that several state-of-the-art attacks can be detected with surprising ease. Second, to address this gap, we introduce Phantasia, a context-adaptive backdoor attack that dynamically aligns its poisoned outputs with the semantics of each input. Instead of producing static poisoned patterns, Phantasia encourages models to generate contextually coherent yet malicious responses that remain plausible, thereby significantly improving stealth and adaptability. Extensive experiments across diverse VLM architectures reveal that Phantasia achieves state-of-the-art attack success rates while maintaining benign performance under various defensive settings.