mlsanityData to 14 Oct 2026

Taxonomy, Opportunities, and Challenges of Representation Engineering for Large Language Models

arXiv 2502.19649 ↗First posted 27 Feb 2025cs.LG cs.CLcited by 34

On the map: Mechanistic interpretability → Knowledge and representations inside LLMs

Abstract (arXiv)

Representation Engineering (RepE) is a novel paradigm for controlling the behavior of LLMs. Unlike traditional approaches that modify inputs or fine-tune the model, RepE directly manipulates the model's internal representations. As a result, it may offer more effective, interpretable, data-efficient, and flexible control over models' behavior. We present the first comprehensive survey of RepE for LLMs, reviewing the rapidly growing literature to address key questions: What RepE methods exist and how do they differ? For what concepts and problems has RepE been applied? What are the strengths and weaknesses of RepE compared to other methods? To answer these, we propose a unified framework describing RepE as a pipeline comprising representation identification, operationalization, and control. We posit that while RepE methods offer significant potential, challenges remain, including managing multiple concepts, ensuring reliability, and preserving models' performance. Towards improving RepE, we identify opportunities for experimental and methodological improvements and construct a guide for best practices.

Key facts. Cited by 34 papers in the map (data through 14 Oct 2026). Placed in Knowledge and representations inside LLMs, one of 13 directions in Mechanistic interpretability. First posted on arXiv 27 Feb 2025.

Builds on

0 of 0 references.

No references here yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 11; 2026: 23.

34 papers.

Sort:

By area:

  1. 2025SEAL: Steerable Reasoning Calibration of Large Language Models for Freecited by 68
  2. 2025The Hawthorne Effect in Reasoning Models: Evaluating and Steering Test Awarenesscited by 30
  3. 2026ODESteer: A Unified ODE-Based Steering Framework for LLM Alignmentcited by 19
  4. 2025Evaluation Awareness Scales Predictably in Open-Weights Large Language Modelscited by 14
  5. 2026Steering Vector Fields for Context-Aware Inference-Time Control in Large Language Modelscited by 13
  6. 2025KV Cache Steering for Controlling Frozen LLMscited by 11
  7. 2025A Representation Engineering Perspective on the Effectiveness of Multi-Turn Jailbreakscited by 10
  8. 2025EasyEdit2: An Easy-to-use Steering Framework for Editing Large Language Modelscited by 7
  9. 2026Steer2Adapt: Dynamically Composing Steering Vectors Elicits Efficient Adaptation of LLMscited by 7
  10. 2026Large Language Models Exhibit Normative Conformitycited by 6
  11. 2026Prompt-Activation Duality: Improving Activation Steering via Attention-Level Interventionscited by 6
  12. 2025The Shape of Adversarial Influence: Characterizing LLM Latent Spaces with Persistent Homologycited by 5
22 more
  1. 2025Neuron-Guided Interpretation of Code LLMs: Where, Why, and How?cited by 5
  2. 2026GAVEL: Towards Rule-Based Safety Through Activation Monitoringcited by 5
  3. 2025A Mixture of Linear Corrections Generates Secure Codecited by 4
  4. 2026Towards Understanding Steering Strengthcited by 3
  5. 2026Knowing without Acting: The Disentangled Geometry of Safety Mechanisms in Large Language Modelscited by 2
  6. 2026Decomposing how prompting steers behaviorcited by 2
  7. 2025Continuous sentiment scores for literary and multilingual contextscited by 1
  8. 2026To Defend Against Cyber Attacks, We Must Teach AI Agents to Hackcited by 1
  9. 2026Fast and Accurate Probing of In-Training LLMs' Downstream Performancescited by 1
  10. 2026RefusalGuard: Geometry-Preserving Fine-Tuning for Safety in LLMscited by 1
  11. 2026High-Dimensional Random Projection for Activation Steering in Language Modelscited by 1
  12. 2026Disentangling Steering Vectorscited by 1
  13. 2025Beyond Hidden-Layer Manipulation: Semantically-Aware Logit Interventions for Debiasing LLMscited by 0
  14. 2026Temporal Preference Concepts and their Functions in a Large Language Modelcited by 0
  15. 2026Beyond a Global Norm: Personalizing Toxicity Sensitivity in Language Models Without Retrainingcited by 0
  16. 2026No Single Neuron of Failure: Distributed Safety Alignment Against White-Box Attackscited by 0
  17. 2026Transformation Laws in Neural Representations: Structure, Realisability, and Constructioncited by 0
  18. 2026Deep Noir: Autonomous Steering Discovery via Architectural Chronometry in Transformer Modelscited by 0
  19. 2026"As a Language Model...": Chat Template Switches LLM Self-Referential Voice and Activation Steering Reproduces Itcited by 0
  20. 2026On the Steering Dimensionality of Refusal in Language Modelscited by 0
  21. 2026Component and Dimension Sparsity in Transformer Refusal Mechanismscited by 0
  22. 2026Model vs. User: Whose Values Prevail? Inference-Time Value Alignment in Multi-Turn Conversationscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Representation Engineering for Large-Language Models: Survey and Research Challenges
  2. 2025Why Representation Engineering Works: A Theoretical and Empirical Study in Vision-Language Models
  3. 2026RepBench: Compiling Benchmarks into Capability Representations for Large Language Models
  4. 2023Aligning Large Language Models with Human Preferences through Representation Engineering
  5. 2024A Timeline and Analysis for Representation Plasticity in Large Language Models
  6. 2024Adversarial Representation Engineering: A General Model Editing Framework for Large Language Models
  7. 2025LMR-BENCH: Evaluating LLM Agent's Ability on Reproducing Language Modeling Research
  8. 2025SR-LLM: Rethinking the Structured Representation in Large Language Model