mlsanityData to 14 Oct 2026

Adaptive Probe-based Steering for Robust LLM Jailbreaking

arXiv 2605.20286 ↗First posted 19 May 2026cs.CR cs.LGcited by 0

On the map: LLM safety and security → Jailbreaking

Abstract (arXiv)

Recent work has demonstrated the potential of contrastive steering for jailbreaking Large Language Models (LLMs). However, existing methods rely on limited and inherently biased contrastive prompts and require laborious manual tuning of steering strength, limiting their robustness and effectiveness. In this paper, we leverage the idea of model extraction to guide the learned steering vectors to approximate the ideal one and propose tuning the steering strength adaptively based on contrastive activations' statistics. Experiments demonstrate that our method notably improves the effectiveness and robustness of probe-based steering, without any extra contrastive prompts or laborious manual tuning. Being an attack paper, this paper focuses on revealing the breakdown of fortified LLMs, raising the average harmfulness score from 6% to 70%. Our code is available at https://github.com/fhdnskfbeuv/adaptiveSteering.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Jailbreaking, one of 12 directions in LLM safety and security. Builds on 34 of its 48 references that are in the map. First posted on arXiv 19 May 2026.

Builds on

34 of 48 references.

Closest ancestors · 11

Same direction, newest first

  1. 2025The Attacker Moves Second: Stronger Adaptive Attacks Bypass Defenses Against Llm Jailbreaks and Prompt Injections10 Oct
  2. 2024AdvPrefix: An Objective for Nuanced LLM Jailbreaks13 Dec
  3. 2024Tamper-Resistant Safeguards for Open-Weight LLMs1 Aug
  4. 2024Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMs22 Jul
  5. 2024Don't Say No: Jailbreaking LLM by Suppressing Refusal25 Apr
  6. 2024AmpleGCG: Learning a Universal and Transferable Generative Model of Adversarial Suffixes for Jailbreaking Both Open and Closed LLMs11 Apr
  7. 2024A StrongREJECT for Empty Jailbreaks15 Feb
  8. 2023Tree of Attacks: Jailbreaking Black-Box LLMs Automatically4 Dec
  9. 2023Jailbreaking Black Box Large Language Models in Twenty Queries12 Oct
  10. 2023AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models3 Oct
  11. 2023Universal and Transferable Adversarial Attacks on Aligned Language Models27 Jul

Other references · 23

  1. 2017Attention Is All You Needcited by 60,361
  2. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  3. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  4. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,720
  5. 2020Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attackscited by 1,451
  6. 2024HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusalcited by 1,429
  7. 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,336
  8. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,268
  9. 2024Refusal in Language Models Is Mediated by a Single Directioncited by 1,032
  10. 2016Stealing Machine Learning Models via Prediction APIscited by 619
  11. 2024Safety Alignment Should Be Made More Than Just a Few Tokens Deepcited by 472
  12. 2023FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Promptscited by 398
11 more
  1. 2025Persona Vectors: Monitoring and Controlling Character Traits in Language Modelscited by 345
  2. 2024Improving Alignment and Robustness with Circuit Breakerscited by 322
  3. 2023Are aligned neural networks adversarially aligned?cited by 315
  4. 2024Personalized Steering of Large Language Models: Versatile Steering Vectors Through Bi-directional Preference Optimizationcited by 118
  5. 2024Refuse Whenever You Feel Unsafe: Improving Safety in LLMs via Decoupled Refusal Trainingcited by 69
  6. 2024Uncovering Safety Risks of Large Language Models through Concept Activation Vectorcited by 67
  7. 2025Trading Inference-Time Compute for Adversarial Robustnesscited by 61
  8. 2025Angular Steering: Behavior Control via Rotation in Activation Spacecited by 53
  9. 2025Representation Bending for Large Language Model Safetycited by 30
  10. 2025To Steer or Not to Steer? Mechanistic Error Reduction with Abstention for Language Modelscited by 29
  11. 2025Does More Inference-Time Compute Really Help Robustness?cited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025AdaSteer: Your Aligned LLM is Inherently an Adaptive Jailbreak Defender
  2. 2024Defending Large Language Models Against Jailbreak Attacks via Layer-specific Editing
  3. 2026Jailbreaking the Matrix: Nullspace Steering for Controlled Model Subversion
  4. 2025Guiding not Forcing: Enhancing the Transferability of Jailbreaking Attacks on LLMs via Removing Superfluous Constraints
  5. 2024Harnessing Task Overload for Scalable Jailbreak Attacks on Large Language Models
  6. 2025Rewrite to Jailbreak: Discover Learnable and Transferable Implicit Harmfulness Instruction
  7. 2025Formalization Driven LLM Prompt Jailbreaking via Reinforcement Learning
  8. 2026Sparse Autoencoders are Capable LLM Jailbreak Mitigators