mlsanityData to 7 Oct 2026

Moving the Safety Barrier: Dynamic Routing Adaptive Alignment Against White-Box Attacks

arXiv 2608.02674 ↗First posted 2 Aug 2026cs.CR cs.AIcited by 0

On the map: LLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks

also in 1 more
Abstract (arXiv)

With the widespread deployment of large foundation models (LFMs) in open environments, safety threats are shifting from black-box jailbreaks toward white-box attacks that directly identify and disrupt internal safety neurons or routes. However, existing safety defenses often rely on static safety units or fixed refusal pathways, leaving models highly vulnerable to targeted route-level white-box attacks. For that, we propose dynamic routing adaptive alignment (DRAA), a framework that introduces dynamic compensatory routes to preserve robust refusal behavior when the safety route is compromised. Specifically, we first identify and localize the model's safety route by contrasting internal activations between safe and unsafe calibration samples. DRAA then masks this safety route to induce causal failure cases and selectively mines the resulting defense failures, thereby constructing failure-aware preference pairs. Extensive experiments demonstrate that DRAA effectively restructures the underlying pathway dependence of model safety, substantially improving robustness against route-level white-box attacks, while preserving general utility.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Safety alignment and jailbreak attacks, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 33 of its 42 references that are in the map. First posted on arXiv 2 Aug 2026.

Builds on

33 of 42 references.

Closest ancestors · 8

Same direction, newest first

  1. 2026SafeNeuron: Neuron-Level Safety Alignment for Large Language Models12 Feb
  2. 2025NeuronTune: Fine-Grained Neuron Modulation for Balanced Safety-Utility Alignment in LLMs13 Aug
  3. 2025NeuRel-Attack: Neuron Relearning for Safety Disalignment in Large Language Models29 Apr
  4. 2024Safety Layers in Aligned Large Language Models: The Key to LLM Security30 Aug
  5. 2024PKU-SafeRLHF: Towards Multi-Level Safety Alignment for LLMs with Human Preference20 Jun
  6. 2024SafeInfer: Context Adaptive Decoding Time Safety Alignment for Large Language Models18 Jun
  7. 2024Language Models are Homer Simpson! Safety Re-Alignment of Fine-tuned Language Models through Task Arithmetic19 Feb
  8. 2023AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models3 Oct

Other references · 25

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  3. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  4. 2018Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challengecited by 4,710
  5. 2023Improved Baselines with Visual Instruction Tuningcited by 4,362
  6. 2021Finetuned Language Models Are Zero-Shot Learnerscited by 3,747
  7. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,919
  8. 2023Sparse Autoencoders Find Highly Interpretable Features in Language Modelscited by 1,359
  9. 2023Inference-Time Intervention: Eliciting Truthful Answers from a Language Modelcited by 1,259
  10. 2024Refusal in Language Models Is Mediated by a Single Directioncited by 960
  11. 2023A Simple and Effective Pruning Approach for Large Language Modelscited by 776
  12. 2021Datasets: A Community Library for Natural Language Processingcited by 526
13 more
  1. 2025Phi-4-Mini Technical Report: Compact yet Powerful Multimodal Language Models via Mixture-of-LoRAscited by 491
  2. 2023XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Modelscited by 477
  3. 2023Red-Teaming Large Language Models using Chain of Utterances for Safety-Alignmentcited by 213
  4. 2024Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMscited by 158
  5. 2024Shaping the Safety Boundaries: Understanding and Defending Against Jailbreaks in Large Language Modelscited by 23
  6. 2026DNA: Uncovering Universal Latent Forgery Knowledgecited by 10
  7. 2025SafePTR: Token-Level Jailbreak Defense in Multimodal LLMs via Prune-then-Restore Mechanismcited by 9
  8. 2025Where Culture Fades: Revealing the Cultural Gap in Text-to-Image Generationcited by 8
  9. 2026TraceRouter: Robust Safety for Large Foundation Models via Path-Level Interventioncited by 8
  10. 2025Unraveling LLM Jailbreaks Through Safety Knowledge Neuronscited by 7
  11. 2026OrthoEraser: Coupled-Neuron Orthogonal Projection for Concept Erasurecited by 3
  12. 2026Stable Attention Response for Reliable Precipitation Nowcastingcited by 2
  13. 2026McCast: Memory-Guided Latent Drift Correction for Long-Horizon Precipitation Nowcastingcited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026When Safety Routing Breaks: Understanding Alignment Fragility under Benign Fine-Tuning
  2. 2026No Single Neuron of Failure: Distributed Safety Alignment Against White-Box Attacks
  3. 2026RASA: Routing-Aware Safety Alignment for Mixture-of-Experts Models
  4. 2025Differentiated Directional Intervention A Framework for Evading LLM Safety Alignment
  5. 2026From Refusal Geometry to Safety Geometry: Harmfulness--Refusal Coupling under Dynamic Adversarial Fine-Tuning
  6. 2026HARC: Coupling Harmfulness and Refusal Directions for Robust Safety Alignment
  7. 2026Dynamic Adversarial Fine-Tuning Reorganizes Refusal Geometry
  8. 2026The Geometry of Refusal: Linear Instability in Safety-Aligned LLMs