# LLM Misalignment via Adversarial RLHF Platforms

arXiv: https://arxiv.org/abs/2503.03039 · HTML: https://mlsanity.com/paper/2503.03039/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Safety alignment and jailbreak attacks, one of 10 directions in LLM alignment / RLHF / preference optimisation.
- Builds on 46 of its 64 references that are in the map.
- First posted on arXiv 4 Mar 2025.

## Placed in

- LLM alignment / RLHF / preference optimisation / [Safety alignment and jailbreak attacks](https://mlsanity.com/alignment/safety-alignment-and-jailbreak-attacks/)

## Builds on (most cited first, 30 of 46)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [Fine-Tuning Language Models from Human Preferences](https://mlsanity.com/paper/1909.08593/) (2019)
- [RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models](https://mlsanity.com/paper/2009.11462/) (2020)
- [Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!](https://mlsanity.com/paper/2310.03693/) (2023)
- [DailyDialog: A Manually Labelled Multi-turn Dialogue Dataset](https://mlsanity.com/paper/1710.03957/) (2017)
- [Improving alignment of dialogue agents via targeted human judgements](https://mlsanity.com/paper/2209.14375/) (2022)
- [RRHF: Rank Responses to Align Language Models with Human Feedback without tears](https://mlsanity.com/paper/2304.05302/) (2023)
- [ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detection](https://mlsanity.com/paper/2203.09509/) (2022)
- [Weight Poisoning Attacks on Pre-trained Models](https://mlsanity.com/paper/2004.06660/) (2020)
- [Certified Defenses for Data Poisoning Attacks](https://mlsanity.com/paper/1706.03691/) (2017)
- [Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models](https://mlsanity.com/paper/2310.02949/) (2023)
- [Data Poisoning Attacks Against Federated Learning Systems](https://mlsanity.com/paper/2007.08432/) (2020)
- [Poisoning Language Models During Instruction Tuning](https://mlsanity.com/paper/2305.00944/) (2023)
- [Is Reinforcement Learning (Not) for Natural Language Processing: Benchmarks, Baselines, and Building Blocks for Natural Language Policy Optimization](https://mlsanity.com/paper/2210.01241/) (2022)
- [Manipulating Machine Learning: Poisoning Attacks and Countermeasures for Regression Learning](https://mlsanity.com/paper/1804.00308/) (2018)
- [Witches' Brew: Industrial Scale Data Poisoning via Gradient Matching](https://mlsanity.com/paper/2009.02276/) (2020)
- [Backdooring Instruction-Tuned Large Language Models with Virtual Prompt Injection](https://mlsanity.com/paper/2307.16888/) (2023)
- [Universal Jailbreak Backdoors from Poisoned Human Feedback](https://mlsanity.com/paper/2311.14455/) (2023)
- [What is in Your Safe Data? Identifying Benign Data that Breaks Safety](https://mlsanity.com/paper/2404.01099/) (2024)
- [On the Exploitability of Instruction Tuning](https://mlsanity.com/paper/2306.17194/) (2023)
- [Certified Robustness to Label-Flipping Attacks via Randomized Smoothing](https://mlsanity.com/paper/2002.03018/) (2020)
- [RLHF Deciphered: A Critical Analysis of Reinforcement Learning from Human Feedback for LLMs](https://mlsanity.com/paper/2404.08555/) (2024)
- [Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey](https://mlsanity.com/paper/2409.18169/) (2024)
- [BadGPT: Exploring Security Vulnerabilities of ChatGPT via Backdoor Attacks to InstructGPT](https://mlsanity.com/paper/2304.12298/) (2023)
- [Label Sanitization against Label Flipping Poisoning Attacks](https://mlsanity.com/paper/1803.00992/) (2018)
- [Generative Poisoning Attack Method Against Neural Networks](https://mlsanity.com/paper/1703.01340/) (2017)
- [From Lists to Emojis: How Format Bias Affects Model Alignment](https://mlsanity.com/paper/2409.11704/) (2024)

## Built on by (most cited first, 5 of 5)

- [A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deployment](https://mlsanity.com/paper/2504.15585/) (2025)
- [LLM Hypnosis: Exploiting User Feedback for Unauthorized Knowledge Injection to All Users](https://mlsanity.com/paper/2507.02850/) (2025)
- [A Jointly Efficient and Optimal Algorithm for Heteroskedastic Generalized Linear Bandits with Adversarial Corruptions](https://mlsanity.com/paper/2602.10971/) (2026)
- [Discovering Implicit Large Language Model Alignment Objectives](https://mlsanity.com/paper/2602.15338/) (2026)
- [The Cognitive Penalty: Ablating System 1 and System 2 Reasoning in Edge-Native SLMs for Decentralized Consensus](https://mlsanity.com/paper/2604.16913/) (2026)

## Abstract (arXiv)

Reinforcement learning has shown remarkable performance in aligning language models with human preferences, leading to the rise of attention towards developing RLHF platforms. These platforms enable users to fine-tune models without requiring any expertise in developing complex machine learning algorithms. While these platforms offer useful features such as reward modeling and RLHF fine-tuning, their security and reliability remain largely unexplored. Given the growing adoption of RLHF and open-source RLHF frameworks, we investigate the trustworthiness of these systems and their potential impact on behavior of LLMs. In this paper, we present an attack targeting publicly available RLHF tools. In our proposed attack, an adversarial RLHF platform corrupts the LLM alignment process by selectively manipulating data samples in the preference dataset. In this scenario, when a user's task aligns with the attacker's objective, the platform manipulates a subset of the preference dataset that contains samples related to the attacker's target. This manipulation results in a corrupted reward model, which ultimately leads to the misalignment of the language model. Our results demonstrate that such an attack can effectively steer LLMs toward undesirable behaviors within the targeted domains. Our work highlights the critical need to explore the vulnerabilities of RLHF platforms and their potential to cause misalignment in LLMs during the RLHF fine-tuning process.