# Red Teaming Multimodal Language Models: Evaluating Harm Across Prompt Modalities and Models

arXiv: https://arxiv.org/abs/2509.15478 · HTML: https://mlsanity.com/paper/2509.15478/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 14 of its 30 references that are in the map.
- First posted on arXiv 18 Sep 2025.

## 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 / [Prompt injection and agent security](https://mlsanity.com/safety/prompt-injection-and-agent-security/)

## Builds on (most cited first, 14 of 14)

- [RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models](https://mlsanity.com/paper/2009.11462/) (2020)
- [Ethical and social risks of harm from Language Models](https://mlsanity.com/paper/2112.04359/) (2021)
- [How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMs](https://mlsanity.com/paper/2401.06373/) (2024)
- [Survey of Vulnerabilities in Large Language Models Revealed by Adversarial Attacks](https://mlsanity.com/paper/2310.10844/) (2023)
- [The Values Encoded in Machine Learning Research](https://mlsanity.com/paper/2106.15590/) (2021)
- [Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models](https://mlsanity.com/paper/2403.09792/) (2024)
- [Jailbreaking Attack against Multimodal Large Language Model](https://mlsanity.com/paper/2402.02309/) (2024)
- [Generative Language Models Exhibit Social Identity Biases](https://mlsanity.com/paper/2310.15819/) (2023)
- [Query-Based Adversarial Prompt Generation](https://mlsanity.com/paper/2402.12329/) (2024)
- [Trading Inference-Time Compute for Adversarial Robustness](https://mlsanity.com/paper/2501.18841/) (2025)
- [Adversarial Attacks and Defenses in Large Language Models: Old and New Threats](https://mlsanity.com/paper/2310.19737/) (2023)
- [Realistic Evaluation of Toxicity in Large Language Models](https://mlsanity.com/paper/2405.10659/) (2024)
- [A Prompting-based Approach for Adversarial Example Generation and Robustness Enhancement](https://mlsanity.com/paper/2203.10714/) (2022)
- [Beyond Perception: Evaluating Abstract Visual Reasoning through Multi-Stage Task](https://mlsanity.com/paper/2505.21850/) (2025)

## Built on by (most cited first, 1 of 1)

- [Same Model, Different Weakness: How Language and Modality Reshape the Jailbreak Attack Surface in Frontier MLLMs](https://mlsanity.com/paper/2605.23157/) (2026)

## Abstract (arXiv)

Multimodal large language models (MLLMs) are increasingly used in real world applications, yet their safety under adversarial conditions remains underexplored. This study evaluates the harmlessness of four leading MLLMs (GPT-4o, Claude Sonnet 3.5, Pixtral 12B, and Qwen VL Plus) when exposed to adversarial prompts across text-only and multimodal formats. A team of 26 red teamers generated 726 prompts targeting three harm categories: illegal activity, disinformation, and unethical behaviour. These prompts were submitted to each model, and 17 annotators rated 2,904 model outputs for harmfulness using a 5-point scale. Results show significant differences in vulnerability across models and modalities. Pixtral 12B exhibited the highest rate of harmful responses (~62%), while Claude Sonnet 3.5 was the most resistant (~10%). Contrary to expectations, text-only prompts were slightly more effective at bypassing safety mechanisms than multimodal ones. Statistical analysis confirmed that both model type and input modality were significant predictors of harmfulness. These findings underscore the urgent need for robust, multimodal safety benchmarks as MLLMs are deployed more widely.