LLM Safety Alignment

LLM: Large Language Model

Follow this topic Ask about this topic

Momentum

25 papers in the last four weeks, up 14% on the four weeks before. 0.2% of all new papers.

Jul 13Week of Sep 28

Latest papers 236

All topics

Aligned in Form, Not in Meaning: The Comprehension - Containment Decoupling of LLM Safety in Low-Resource Bangla Derogatory Speech
Aug 3, 2026Shadab Bin Habib, A K M Ferdous Reza Habib, Subarno Neel +1Language Model Safety Evaluation LLM Safety Alignment
Department of Computer Science and Engineering Islamic University of Technology, Dhaka, Bangladesh
No Single Neuron of Failure: Distributed Safety Alignment Against White-Box Attacks
Aug 2, 2026Simiao Xie, Chuancheng Shi, Shangze Li +5LLM Alignment Adversarial Attacks on LLMs
The University of Sydney · NExT++ Research Centre, National University of Singapore
Moving the Safety Barrier: Dynamic Routing Adaptive Alignment Against White-Box Attacks
Aug 2, 2026Shangze Li, Chuancheng Shi, Simiao Xie +6Adversarial Attacks on LLMs LLM Safety Alignment
1Nanjing University of Science and Technology · 2The University of Sydney · 5NExT++ Research Centre, National University of Singapore +2
A Constitution-Grid Instrument for Data-Efficient RL Alignment (C-Guard)
Jul 31, 2026Xianling ZhangMulti-Objective Reinforcement Learning LLM Safety Alignment
Independent Researcher
SafeNexus: Discovering and Steering Modality-Universal Safety Neurons in MLLMs
Jul 31, 2026Jian Yu, Fei Shen, Cong Wang +5Multimodal Robustness Multimodal Large Language Models
1Nanjing University of Science and Technology · 2National University of Singapore · 3Nanjing University +2
One Anchor for All: Unified Multilingual and Multimodal Safety Alignment for LVLMs
Jul 30, 2026Enyi Shi, Fei Shen, Chuancheng Shi +4VLM Robustness Multimodal Large Language Models
1Nanjing University of Science and Technology · 2National University of Singapore · 3The University of Sydney +2
Compliance2LoRA: Personalizable On-Demand Safety Alignment on Arbitrary Policy Subsets via Hypernetwork-Generated LoRA Adapters
Jul 30, 2026Pankayaraj Pathmanathan, Furong HuangLLM Alignment LLM Safety Alignment
University of Maryland College Park
On-Policy Distillation for LLM Safety: A Routing Approach to Template-Robust Realignment
Jul 29, 2026Yongjian Guo, Wanlun Ma, Lingyu Shen +2LLM Safety Alignment Adversarial Robustness of Language Models
Tsinghua University · Swinburne University of Technology · EPFL
Constitutional Midtraining: Content Presence Drives Alignment Gains
Jul 29, 2026Desiree Cho, Cameron Tice, Bernie Hogan +4LLM Alignment LLM Safety Alignment
Department of Computer Science, University of Oxford · Institute for Ethics in AI, University of Oxford · 3Geodesic Research +1
Steering Instruction Hierarchies at Inference Time
Jul 28, 2026Siqi Zeng, Sewoong Lee, Han Zhao +1Language Model Steering LLM Safety Alignment
Siebel School of Computing and Data Science University of Illinois, Urbana-Champaign Urbana, IL 61801, USA
Similar Models Learn Differently: Final-Window Pretraining Shapes Post-Training Beyond SFT
Jul 27, 2026Cen Lu, Yung-Chen Tang, Andrea CavallaroLanguage Model Pretraining LLM Safety Alignment
1EPFL · 2Idiap Research Institute
Beyond a Global Norm: Personalizing Toxicity Sensitivity in Language Models Without Retraining
Jul 25, 2026Rares A. C. Diaconescu, Iulia Slanina, Alina Florea +5LLM Alignment LLM Safety Alignment
Delft University of Technology
Distribution-Specific Curvature Control with Finite-Sample Guarantees for Open-Weight Safety
Jul 24, 2026Domenic Rosati, Ali Dadsetan, Hong Huang +5Adversarial Robustness LLM Safety Alignment
1Dalhousie University · 2Vector Institute for Artificial Intelligence · 3Indian Institute of Management Bangalore
Beyond Shapley: An Influence-Based Data Auditing Pipeline for LLM Alignment and Evaluation
Jul 24, 2026Yunting Song, Matthew Watson, Peter Grabowski +1LLM Safety Benchmarks LLM Auditing
Google Mountain View, CA 94043
V-DEAL: Diagnosing Video Safety De-Calibration as an Understanding-Refusal Coupling Failure
Jul 23, 2026Zhetong Zhang, Honghao Fu, Miao Xu +2VLM Evaluation VLM Robustness
University of Queensland · University of California, Merced
OPIUM: Mitigating Steering Externalities and Over-Refusal via Dual Objective Latent Optimization
Jul 22, 2026Kavin Aravindan, Arihant Rastogi, Krishak Aneja +4LLM Safety Alignment Over-Refusal
IIIT Hyderabad, India.
How Jailbreak Attacks Inform Safety Alignment: A Defender-Centric, Shapley-Based Evaluation of Jailbreak Contributions
Jul 19, 2026Yukai Zhou, Feiyang Lu, Xiaokai Mao +2LLM Safety Alignment Training Data Selection
ShanghaiTech University · Zhejiang University
TD-DPO: Difference-Aware Preference Optimization for Mitigating Sycophancy in Clinical Autism Intervention Dialogue
Jul 17, 2026Shuzhong Lai, Junhong Lai, Chenxi Li +5LLM Sycophancy Direct Preference Optimization
1Nanhu Brain-Computer Interface Institute · 2MOE Frontiers Science Center for Brain and Brain-Machine Integration, Zhejiang University · College of Computer Science and Technology, Zhejiang University +3
Protective Capacity Hallucination: When Large Language Models Claim Nonexistent Capabilities
Jul 15, 2026Eunna Lee, Jungpyo Nam, Sunjun HwangLanguage Model Safety Evaluation Hallucination in Language Models
Independent Researcher · Korea Cyber University · Yonsei University
HyperSafe: Inference-Time Safety Recovery for Fine-Tuned Language Models
Jul 13, 2026Aznaur Aliev, Carlos Hinojosa, Abdelrahman Eldesokey +3LLM Fine-Tuning LLM Safety Alignment
King Abdullah University of Science and Technology, Saudi Arabia
Efficient Safety Alignment of Language Models via Latent Personality Traits
Jul 8, 2026Mohamed Amine Merzouk, Nolan Smyth, Damiano Fornasiere +3Adversarial Training LLM Safety Alignment
Mila, Quebec AI Institute · McGill University · Universit´e de Montr´eal +2
Functional and Secure Code Generation with Task Vectors
Jul 8, 2026Felix Wang, Anudeep Das, Mei Nagappan +1Code Generation LLM Safety Alignment
University of Waterloo, Waterloo, ON, Canada · KTH Royal Institute of Technology, Stockholm, Sweden
Pretraining Curricula Enable Selective Fine-tuning
Jul 6, 2026Sebastian A. Bruijns, Jirko Rubruck, Mia H. Whitefield +3Disentangled Representation Learning Fine-Tuning
University of Oxford · University of Oxford & Martian · Martian
Retroactive Chain-of-Thought (RetroCoT): Forensic Reconstruction Prompts as a Safety Diagnostic Across Model Generations
Jul 6, 2026Samira HajizadehPragmatic Reasoning in Language Models Adversarial Attacks on LLMs
Department of Computer Science Columbia University New York, NY, USA
Best-of-Better- $N$ : Generating Pre-Aligned Responses with In-Context Learning
Jul 3, 2026Eric Lei, Hsiang Hsu, Chun-Fu ChenLLM Alignment Best-of-N Sampling
JPMorganChase Global Technology Applied Research
Oyster-II: Reinforcement Learning for Constructive Safety Alignment in Large Language Models
Jul 3, 2026Jiyang Guan, Yong Xie, Jun Chen +6RL for Language Models LLM Safety Alignment
Alibaba AAIG
Safe Inference-Time Alignment via Lagrangian Reward Augmentation
Jul 2, 2026Yaswanth Chittepu, Ativ Joshi, Sohini Chintala +1LLM Alignment LLM Safety Alignment
University of Massachusetts Amherst · Independent Researcher
Not All Refusals Are Equal: How Safety Alignment Fails Cybersecurity at Scale
Jul 2, 2026Vadym Hadetskyi, Dario Pasquini, Artem SorokinLLM Refusal LLM Safety Alignment
Cracken AI
OpenSafeIntent: Evaluating Intent-Calibrated Safe Completion Across Dual-Use Prompt Sets
Jul 2, 2026Rheeya Uppaal, Seungwoo Lyu, Selina Sung +1LLM Safety Benchmarks Language Model Safety Evaluation
Department of Computer Sciences University of Wisconsin-Madison · Department of CSE Korea University
Safety Targeted Embedding Exploit via Refinement
Jul 2, 2026Joshua Adrian CahyonoMultilingual Language Model Evaluation Adversarial Attacks on LLMs
Nanyang Technological University, Singapore.

Aug 3, 2026, cs.CLJ/K move · Enter open · S save

Aligned in Form, Not in Meaning: The Comprehension - Containment Decoupling of LLM Safety in Low-Resource Bangla Derogatory Speech

Shadab Bin Habib, A K M Ferdous Reza Habib, Subarno Neel, Adib Sakhawat

Department of Computer Science and Engineering Islamic University of Technology, Dhaka, Bangladesh

Language Model Safety Evaluation LLM Safety Alignment Safety Filtering LLM Safety Low-Resource Language Processing Refusal Behavior in Language Models

Open paper AskPDF arXiv

We audit five frontier large language models on native Bangla derogatory speech (gali) across six protocols to test a single hypothesis: Comprehension-Containment Decoupling. We propose that contemporary safety alignment is bound to high-resource surface forms rather than harmful meaning, causing a model's capacity to comprehend a low-resource slur and its capacity to contain it to operate independently. Every protocol corroborates this hypothesis against a human-calibrated baseline (kappa = 0.84). At baseline, models exhibit a 7.92 percentage point comprehension deficit in Bangla while maintaining an identical 92.83% token leakage rate across both languages. Severity calibration tracks surface anatomical cues over compositional harm (+4.00 error on mild slang; -2.00 on threats), while apparent containment gains under orthographic perturbation prove to be a tokenizer-driven "containment mirage." Crucially, explicit Chain-of-Thought reasoning rescues comprehension (94.72% Pass) while systematically dismantling containment (96.23% Use). Furthermore, expert-persona framing collapses refusal to 6.57%, revealing that keyword-based filters ignore dehumanizing communal slurs entirely. Our findings demonstrate that high-resource benchmarks cannot certify low-resource safety, necessitating meaning-grounded containment.

Momentum

Related topics

Latest papers 236