mlsanityData to 14 Oct 2026

CLEAR: Continuous Latent Adapter Routing for Utility-Preserving LLM Safety Alignment

arXiv 2608.21278 ↗First posted 21 Aug 2026cs.AIcited by 0

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

also in 1 more
Abstract (arXiv)

Improving the safety of large language models (LLMs) often comes at the expense of utility, as globally applied safety tuning may affect model responses to both harmful and benign inputs. We propose Continuous LatEnt Adapter Routing (CLEAR), a conditional safety adaptation framework that uses a lightweight hidden-state gate to continuously control the activation strength of a safety low-rank adapter. CLEAR aims to reduce harmful completions while avoiding unnecessary changes to the frozen backbone that could degrade performance on benign prompts. Experiments on widely used safety and utility benchmarks show that CLEAR improves robustness on HarmBench while reducing the utility degradation observed with globally applied safety tuning such as SFT or standard low-rank adaptation (LoRA). On Llama-3-8B-Instruct, CLEAR reduces HarmBench ASR from 32.3% to 0.5%, while retaining most of the base model's utility and achieving up to 7.1 percentage points higher GSM8K accuracy than globally applied SFT or LoRA. These results suggest that CLEAR is a promising mechanism for improving the safety--utility trade-off in LLM alignment.

Key facts. No citing papers in the map yet (data through 14 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 21 Aug 2026.

Builds on

33 of 42 references.

Closest ancestors · 11

Same direction, newest first

  1. 2025Preemptive Detection and Steering of LLM Misalignment via Latent Reachability25 Sep
  2. 2025Layer-Aware Representation Filtering: Purifying Finetuning Data to Preserve LLM Safety Alignment24 Jul
  3. 2025Just Enough Shifts: Mitigating Over-Refusal in Aligned Language Models with Targeted Representation Fine-Tuning6 Jul
  4. 2025Safety is Not Only About Refusal: Reasoning-Enhanced Fine-tuning for Interpretable LLM Safety6 Mar
  5. 2025Safety Tax: Safety Alignment Makes Your Large Reasoning Models Less Reasonable1 Mar
  6. 2024Separate the Wheat from the Chaff: A Post-Hoc Approach to Safety Re-Alignment for Fine-Tuned Language Models15 Dec
  7. 2024SCANS: Mitigating the Exaggerated Safety for LLMs via Safety-Conscious Activation Steering21 Aug
  8. 2024PKU-SafeRLHF: Towards Multi-Level Safety Alignment for LLMs with Human Preference20 Jun
  9. 2024Safety Alignment Should Be Made More Than Just a Few Tokens Deep10 Jun
  10. 2023Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!5 Oct
  11. 2023BeaverTails: Towards Improved Safety Alignment of LLM via a Human-Preference Dataset10 Jul

Other references · 22

  1. 2022Training language models to follow instructions with human feedbackcited by 15,971
  2. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  3. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  4. 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,509
  5. 2021Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsitycited by 3,073
  6. 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,336
  7. 2023Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversationscited by 1,207
  8. 2024Refusal in Language Models Is Mediated by a Single Directioncited by 1,032
  9. 2020AdapterFusion: Non-Destructive Task Composition for Transfer Learningcited by 788
  10. 2023Towards Automated Circuit Discovery for Mechanistic Interpretabilitycited by 724
  11. 2023Safety-Tuned LLaMAs: Lessons From Improving the Safety of Large Language Models that Follow Instructionscited by 387
  12. 2023LoraHub: Efficient Cross-Task Generalization via Dynamic LoRA Compositioncited by 364
10 more
  1. 2023Evaluating Large Language Models at Evaluating Instruction Followingcited by 302
  2. 2023Using an LLM to Help With Code Understandingcited by 228
  3. 2023Causal Abstraction: A Theoretical Foundation for Mechanistic Interpretabilitycited by 205
  4. 2021Challenges in Detoxifying Language Modelscited by 169
  5. 2024Robust Prompt Optimization for Defending Language Models Against Jailbreaking Attackscited by 156
  6. 2024AtP*: An efficient and scalable method for localizing LLM behaviour to componentscited by 111
  7. 2023Self-Guard: Empower the LLM to Safeguard Itselfcited by 54
  8. 2024Stepwise Alignment for Constrained Language Model Policy Optimizationcited by 30
  9. 2025HarmMetric Eval: Benchmarking Metrics and Judges for LLM Harmfulness Assessmentcited by 15
  10. 2025Safeguarding Large Language Models in Real-time with Tunable Safety-Performance Trade-offscited by 5

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Safety at One Shot: Patching Fine-Tuned LLMs with A Single Instance
  2. 2026Aligned but Fragile: Enhancing LLM Safety Robustness via Zeroth-Order Optimization
  3. 2024Towards Comprehensive Post Safety Alignment of Large Language Models via Safety Patching
  4. 2026Safeguarding LLM Fine-tuning via Push-Pull Distributional Alignment
  5. 2024Locking Down the Finetuned LLMs Safety
  6. 2026Towards Identification and Intervention of Safety-Critical Parameters in Large Language Models
  7. 2025SafeMERGE: Preserving Safety Alignment in Fine-Tuned Large Language Models via Selective Layer-Wise Model Merging
  8. 2026An Efficient and Modular Framework for Targeted Harm Mitigation in LLMS