# Feature-Aware Malicious Output Detection and Mitigation

arXiv: https://arxiv.org/abs/2504.09191 · HTML: https://mlsanity.com/paper/2504.09191/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 15 of its 19 references that are in the map.
- First posted on arXiv 12 Apr 2025.

## Builds on (most cited first, 15 of 15)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [Jailbreaking Black Box Large Language Models in Twenty Queries](https://mlsanity.com/paper/2310.08419/) (2023)
- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [Ethical and social risks of harm from Language Models](https://mlsanity.com/paper/2112.04359/) (2021)
- [Baseline Defenses for Adversarial Attacks Against Aligned Language Models](https://mlsanity.com/paper/2309.00614/) (2023)
- [Jailbreaking ChatGPT via Prompt Engineering: An Empirical Study](https://mlsanity.com/paper/2305.13860/) (2023)
- [Detecting Language Model Attacks with Perplexity](https://mlsanity.com/paper/2308.14132/) (2023)
- [LLM Self Defense: By Self Examination, LLMs Know They Are Being Tricked](https://mlsanity.com/paper/2308.07308/) (2023)
- [Multimodal Large Language Models: A Survey](https://mlsanity.com/paper/2311.13165/) (2023)
- [A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity](https://mlsanity.com/paper/2401.01967/) (2024)
- [How Alignment and Jailbreak Work: Explain LLM Safety through Intermediate Hidden States](https://mlsanity.com/paper/2406.05644/) (2024)

## Abstract (arXiv)

The rapid advancement of large language models (LLMs) has brought significant benefits to various domains while introducing substantial risks. Despite being fine-tuned through reinforcement learning, LLMs lack the capability to discern malicious content, limiting their defense against jailbreak. To address these safety concerns, we propose a feature-aware method for harmful response rejection (FMM), which detects the presence of malicious features within the model's feature space and adaptively adjusts the model's rejection mechanism. By employing a simple discriminator, we detect potential malicious traits during the decoding phase. Upon detecting features indicative of toxic tokens, FMM regenerates the current token. By employing activation patching, an additional rejection vector is incorporated during the subsequent token generation, steering the model towards a refusal response. Experimental results demonstrate the effectiveness of our approach across multiple language models and diverse attack techniques, while crucially maintaining the models' standard generation capabilities.