# Representation Bending for Large Language Model Safety

arXiv: https://arxiv.org/abs/2504.01550 · HTML: https://mlsanity.com/paper/2504.01550/

## Key facts

- Cited by 30 papers in the map (data through 14 Oct 2026).
- Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security.
- Builds on 65 of its 90 references that are in the map.
- First posted on arXiv 2 Apr 2025.

## Placed in

- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 30 of 65)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [Deep reinforcement learning from human preferences](https://mlsanity.com/paper/1706.03741/) (2017)
- [Qwen Technical Report](https://mlsanity.com/paper/2309.16609/) (2023)
- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them](https://mlsanity.com/paper/2210.09261/) (2022)
- [Jailbroken: How Does LLM Safety Training Fail?](https://mlsanity.com/paper/2307.02483/) (2023)
- [HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal](https://mlsanity.com/paper/2402.04249/) (2024)
- [Representation Engineering: A Top-Down Approach to AI Transparency](https://mlsanity.com/paper/2310.01405/) (2023)
- [Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!](https://mlsanity.com/paper/2310.03693/) (2023)
- [Steering Llama 2 via Contrastive Activation Addition](https://mlsanity.com/paper/2312.06681/) (2023)
- [Refusal in Language Models Is Mediated by a Single Direction](https://mlsanity.com/paper/2406.11717/) (2024)
- [Steering Language Models With Activation Engineering](https://mlsanity.com/paper/2308.10248/) (2023)
- [Gemma: Open Models Based on Gemini Research and Technology](https://mlsanity.com/paper/2403.08295/) (2024)
- [Enhancing Chat Language Models by Scaling High-quality Instructional Conversations](https://mlsanity.com/paper/2305.14233/) (2023)
- [Eliciting Latent Predictions from Transformers with the Tuned Lens](https://mlsanity.com/paper/2303.08112/) (2023)
- [How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMs](https://mlsanity.com/paper/2401.06373/) (2024)
- [TOFU: A Task of Fictitious Unlearning for LLMs](https://mlsanity.com/paper/2401.06121/) (2024)
- [The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning](https://mlsanity.com/paper/2403.03218/) (2024)
- [Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearning](https://mlsanity.com/paper/2404.05868/) (2024)
- [WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs](https://mlsanity.com/paper/2406.18495/) (2024)
- [Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks](https://mlsanity.com/paper/2404.02151/) (2024)
- [Safety Alignment Should Be Made More Than Just a Few Tokens Deep](https://mlsanity.com/paper/2406.05946/) (2024)

## Built on by (most cited first, 30 of 30)

- [There Is More to Refusal in Large Language Models than a Single Direction](https://mlsanity.com/paper/2602.02132/) (2026)
- [SAFEPATH: Preventing Harmful Reasoning in Chain-of-Thought via Early Alignment](https://mlsanity.com/paper/2505.14667/) (2025)
- [Consistency Training Helps Stop Sycophancy and Jailbreaks](https://mlsanity.com/paper/2510.27062/) (2025)
- [A2D: Any-Order, Any-Step Safety Alignment for Diffusion Language Models](https://mlsanity.com/paper/2509.23286/) (2025)
- [Improving Large Language Model Safety with Contrastive Representation Learning](https://mlsanity.com/paper/2506.11938/) (2025)
- [TRACES: Proactive Safety Auditing for Multi-Turn LLM Agents via Trajectory-State Modeling](https://mlsanity.com/paper/2605.27690/) (2026)
- [ARGUS: Defending Against Multimodal Indirect Prompt Injection via Steering Instruction-Following Behavior](https://mlsanity.com/paper/2512.05745/) (2025)
- [Hallucination reduction with CASAL: Contrastive Activation Steering For Amortized Learning](https://mlsanity.com/paper/2510.02324/) (2025)
- [Turning the Spell Around: Lightweight Alignment Amplification via Rank-One Safety Injection](https://mlsanity.com/paper/2508.20766/) (2025)
- [Adversarial Contrastive Learning for LLM Quantization Attacks](https://mlsanity.com/paper/2601.02680/) (2026)
- [GAVEL: Towards Rule-Based Safety Through Activation Monitoring](https://mlsanity.com/paper/2601.19768/) (2026)
- [Minimizing Collateral Damage in Activation Steering](https://mlsanity.com/paper/2605.01167/) (2026)
- [DUSK: Do Not Unlearn Shared Knowledge](https://mlsanity.com/paper/2505.15209/) (2025)
- [Large Language Models Generate Harmful Responses Using a Distinct Mechanism, Shared Across Harm Types](https://mlsanity.com/paper/2604.09544/) (2026)
- [Knowing without Acting: The Disentangled Geometry of Safety Mechanisms in Large Language Models](https://mlsanity.com/paper/2603.05773/) (2026)
- [ASGuard: Activation-Scaling Guard to Mitigate Targeted Jailbreaking Attack](https://mlsanity.com/paper/2509.25843/) (2025)
- [Understanding and Mitigating Over-refusal for Large Language Models via Representation Intervention](https://mlsanity.com/paper/2511.19009/) (2025)
- [TrojanPraise: Jailbreak LLMs via Benign Fine-Tuning](https://mlsanity.com/paper/2601.12460/) (2026)
- [Beyond Creed: A Non-Identity Safety Condition A Strong Empirical Alternative to Identity Framing in Low-Data LoRA Fine-Tuning](https://mlsanity.com/paper/2603.14723/) (2026)
- [RefusalGuard: Geometry-Preserving Fine-Tuning for Safety in LLMs](https://mlsanity.com/paper/2605.01913/) (2026)
- [DataShield: Safety-degrading Data Filtering for LLM Benign Instruction Fine-Tuning](https://mlsanity.com/paper/2606.00160/) (2026)
- [Decoy Direction Optimization: A Post-Hoc Defense Against LLM Abliteration](https://mlsanity.com/paper/2609.16204/) (2026)
- [Analysing LLM Persona Generation and Fairness Interpretation in Polarised Geopolitical Contexts](https://mlsanity.com/paper/2603.22837/) (2026)
- [Adaptive Probe-based Steering for Robust LLM Jailbreaking](https://mlsanity.com/paper/2605.20286/) (2026)
- [LoMC: Localized Multidirectional Correction for Refusal Suppression in Routed Foundation Models](https://mlsanity.com/paper/2606.13709/) (2026)
- [HARC: Coupling Harmfulness and Refusal Directions for Robust Safety Alignment](https://mlsanity.com/paper/2607.00572/) (2026)
- [Mitigating Scaffolding Collapse in Socratic Tutors via Representation Alignment](https://mlsanity.com/paper/2607.19371/) (2026)
- [SafeTune: A Unified Faithful Library for Auditing and Repairing Safety Drift in Fine-Tuned LLMs](https://mlsanity.com/paper/2609.22153/) (2026)
- [AnchorRep: Defending LLMs Against Cross-Model Adversarial Transfer via Representation Repulsion](https://mlsanity.com/paper/2609.32602/) (2026)
- [How Narrative Wrapping Affects LLM Refusal: A Cross-Language Benchmark and Defense](https://mlsanity.com/paper/2610.11005/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) have emerged as powerful tools, but their inherent safety risks - ranging from harmful content generation to broader societal harms - pose significant challenges. These risks can be amplified by the recent adversarial attacks, fine-tuning vulnerabilities, and the increasing deployment of LLMs in high-stakes environments. Existing safety-enhancing techniques, such as fine-tuning with human feedback or adversarial training, are still vulnerable as they address specific threats and often fail to generalize across unseen attacks, or require manual system-level defenses. This paper introduces RepBend, a novel approach that fundamentally disrupts the representations underlying harmful behaviors in LLMs, offering a scalable solution to enhance (potentially inherent) safety. RepBend brings the idea of activation steering - simple vector arithmetic for steering model's behavior during inference - to loss-based fine-tuning. Through extensive evaluation, RepBend achieves state-of-the-art performance, outperforming prior methods such as Circuit Breaker, RMU, and NPO, with up to 95% reduction in attack success rates across diverse jailbreak benchmarks, all with negligible reduction in model usability and general capabilities.