Steering Llama 2 via Contrastive Activation Addition

Rethinking Refusal Directions in Large Language Models: Alignment-Driven Behavior and Cross-Lingual Variability

Liu, W. T., Xie, W., Wang, Z., Ma, S., & Tong, H. (2026) · Communications in computer and information science · doi:10.1007/978-981-92-3541-4_7

Ethical and Legal Limits of Aligning Large Language Models: The Illusion of Control

Schmitt, V., Ostermann, S., & Solopova, V. (2026) · Artificial intelligence · doi:10.5772/intechopen.1016018

Steering into Weights: Persistent Behavioral Editing for Targeted Harm Categories

Liu, X., Cao, D., Cai, Y., Huang, Y. N., & Liu, G. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-3435-6_12

Engineering Safety Into LLMs: How to Train Safety Constraints as Learned ConceptsPreprint

Roth, R. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6926647

Security-by-Design for LLM-Based Code Generation: Leveraging Internal Representations for Concept-Driven Steering Mechanisms

Wendlinger, M., Kowatsch, D., Böttinger, K., & Sperl, P. (2026) · IEEE European Symposium on Security and Privacy · doi:10.1109/eurosp68448.2026.00049 · Cited by 2

A Geometric Information Bottleneck for Activation Steering

Doan, T., Nguyen, T., & Gupta, S. (2026) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3770855.3818109

ActivationBackdoor: Backdooring Large Language Models in Collaborative Inference via Intermediate Activations

Su, Z., Zhang, M., Zhang, X., Hong, G., You, X., & Yang, M. (2026) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3770855.3818136

HyperEdit: Mitigating Hallucinations of Large Language Models via Hyperbolic Representation Editing

Lin, T., Du, J., Xue, Z., Liang, M., & Tang, R. (2026) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3770855.3818153

Towards principled knowledge editing methods for large language model reasoning

Yao, Y., Qin, J., Zhang, N., Xu, H., Zhu, Y., Yu, Z., Wang, M., Tang, Y., Gu, J.-C., Deng, S., et al. (2026) · Nature Machine Intelligence · doi:10.1038/s42256-026-01276-y

Restoring the Right Stream: Training-Free OOD Robustness for Vision–Language–Action Policies

Yan, Z., & Shen, G. (2026) · Entropy · doi:10.3390/e28090975

Cross-Modal Safety-Component Steering for Audio-Language Model Jailbreak DefensePreprint

Xie, P., & Liu, X. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7417884

When Not to Edit: Reliability Diagnostics for Activation-Space Interventions in Language Models

Osafo, N. Y. A., Doku, R., & Rawat, D. B. (2026) · IEEE International Conference on Information Reuse and Integration (IRI) · doi:10.1109/iri69576.2026.00026

Latent Steering of Emotional Evidence with Sparse Autoencoders in Language Models

Quintero-Sánchez, E. A., López-Ramírez, C., Ramos-Aguilar, R., García-Ramírez, J., & Ramon-Pedrero, L. E. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-37441-7_13

Auditing Abliteration: Normalization Placement, Edit Budget, and Transfer in Single-Direction Refusal Removal

Tucker, C., & León, M. (2026) · IEEE Opportunity Research Scholars Symposium (ORSS) · doi:10.1109/orss71174.2026.11684276

Residual Information Network Editing for Persona Steering in Frozen Language Models

Pratama, A. A. Y., Kliveson, C., Elhan, A., & Hidayaturrahman (2026) · International Conference on Information Management and Technology (ICIMTech) · doi:10.1109/icimtech202671078.2026.11699070

LLM-Lobotomy: A Representation Engineering Framework for LLM Behavioral Steering

Neto, E. B., Ferreira, L. N., Interian, R., & Dias, Z. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-39882-6_39

TriSAIL: A Source-Sensitive Audit Protocol for Symmetric and Asymmetric Transfer in Hidden-State Safety Probes

Chen, G., Xia, Y., Ma, X., & Li, Z. M. (2026) · Symmetry · doi:10.3390/sym18101644

Co-steer: Cross-Modal Collaborative Steering for Jailbreaking MLLMs

Zhu, J., Omari, R. M., Abraham, T., Kim, J., Kaur, A., Le, T., Phung, D., & Ke, Q. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37225-3_14

Fully Unsupervised Cross-Lingual Capabilities Enhancement Based on Language-Specific Neurons

Sun, Q., Liu, Z., Cao, H., Yang, M., & Zhao, T. (2026) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/taslpro.2026.3739186

Decoding Future Semantics from Language Model Hidden States: Rollout Trade-offs and Causal SteeringPreprint

Choi, J. (2026) · Research Square · doi:10.21203/rs.3.rs-10973324/v1

VISOR++ : Visual Input Based Steering for Large Vision Language Models

Balakrishnan, R., & Phute, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37086-0_27

Fairness Guardrails for Autonomous Agents: A Trustworthiness-by-Design Vision from Social Cybersecurity

Baldassarre, M. T., Barletta, V. S., Bavaro, V., & Santos, R. E. S. (2026) · doi:10.1145/3843782.3844661

The Pain–Comfort Lab: Looking Inside a Language Model

Menon, P. (2026) · Zenodo (CERN European Organization for Nuclear Research) · doi:10.5281/zenodo.23228532

Enhancing AI Interpretability Through Localised Architectures

Seet, I., Bozenhard, J., & Ostermann, S. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-41269-0_10

Beyond Attention: Module-Level Refusal Directions in Mamba-Based Hybrid Language Models

Kirilchuk, A., & Malykh, V. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-41546-2_10

BhashaLLM: A QLoRA-Based Framework for Bangla Text Generation and Handwritten Character Recognition

Binaaf, S. A., Arafat, Y., Chaklader, A. S., & Rahman, M. R. (2026) · IEEE International Conference on Intelligent Systems · doi:10.1109/is71416.2026.11727246

Latent Representation of Psychological Constructs in Large Language ModelsPreprint

Chen, J. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7652158

Amplifying Latent Safety in Large Language Models via Gated Activation Steering

Prosvirnin, R., Suslov, A., Soldatov, A., Kutishcheva, A., Litikov, P., & Bashun, V. (2026) · Computation · doi:10.3390/computation14100261

Paired response final divergence signatures of jailbreak vulnerability in large language models

Ismail, H. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-02593-2

Depressive Belief Rigidity in Language Models: Representation and Limits of Activation-Based Control

Xie, S. (2026) · International Artificial Intelligence and Data Processing Symposium (IDAP) · doi:10.1109/idap71869.2026.11727911

HintSteer: Predicting Context Imprints for Inference-Time Steering

Wang, X., Zhao, C., Hu, Y., Li, G., Ou, C., & Zheng, Y. (2026) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/2026.findings-emnlp.656

Inside self-reflection: the geometry in multi-dimensional scoring of metacognitive structure

Suthaharan, P., Corlett, P. R., & Ang, Y. (2026) · BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP (BlackboxNLP) · doi:10.18653/2026.blackboxnlp-1.9

Inoculate or Reflect? Two Training Interventions Under Prompting, Steering, and Patching

Imran, A., & Aaliyan, M. (2026) · BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP (BlackboxNLP) · doi:10.18653/2026.blackboxnlp-1.64

Revisiting Task Vectors: A Structural Analysis of Their Limits

Han, L., Hyeseung, J. I., Jung, S., Lyu, S., & Seo, P. H. (2026) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/2026.findings-emnlp.1606

Geometric Signatures of Post-Training in LLM Reasoning Trajectories: Separating Observability from Transferability

Chen, T., & Cai, Y. (2026) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl.a.819

Steering Large Language Models toward Cross-Cultural Value Alignment via Activation Engineering

Zhang, J., Lin, J., Yang, Y., Zhou, Y., & Hu, S. (2026) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3799682.3840854

Personalized Text Generation with Contrastive Activation Steering

Zhang, J., Liu, Y., Wang, W., Liu, Q., Wu, S., Wang, L., & Chua, T. (2025) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2025.acl-long.353 · Cited by 5

Contrastive Covariance Subspaces for Activation SteeringPreprint

Xenos, A., Plitsis, M., Foteinopoulou, N., Panagakis, Y., & Patras, I. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2610.14025

Steering Speech-Language Models: Training-Free Task Specialization via Contrastive Activation AdditionPreprint

Baroudi, S., Labrak, Y., Melucci, P., Burdisso, S., Motlíček, P., Bredin, H., Ravanelli, M., & Marxer, R. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2610.04683

Contrastive Prompting Enhances Sentence Embeddings in LLMs through Inference-Time Steering

Cheng, Z., Wang, Z. J., Fu, Y., Jiang, Z., Yin, Y., Wang, C., & Gu, Q. (2025) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2025.acl-long.174 · Cited by 4

Scaling laws for activation steering with Llama 2 models and refusal mechanismsPreprint

Ali, S. A. R., Xu, J., Yang, I., Li, J. X., Arslan, A., & Benham, C. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2507.11771 · Cited by 2