# Activation steering methods Area: [Mechanistic interpretability](https://mlsanity.com/mechinterp/index.md). HTML: https://mlsanity.com/mechinterp/activation-steering-methods/ ## Key facts - Activation steering methods: 239 papers on arXiv since 2022 (data through 14 Oct 2026). - 159 of its 239 mapped papers (67%) cite “Steering Llama 2 via Contrastive Activation Addition” (2023, arXiv 2312.06681). - 99 of its 239 mapped papers (41%) cite “Steering Language Models With Activation Engineering” (2023, arXiv 2308.10248). - 28 new papers 24 Sep–14 Oct. ## Most cited here - [Steering Llama 2 via Contrastive Activation Addition](https://mlsanity.com/paper/2312.06681/) (2023), cited by 159 papers here - [Steering Language Models With Activation Engineering](https://mlsanity.com/paper/2308.10248/) (2023), cited by 99 papers here - [Analyzing the Generalization and Reliability of Steering Vectors](https://mlsanity.com/paper/2407.12404/) (2024), cited by 33 papers here - [Extracting Latent Steering Vectors from Pretrained Language Models](https://mlsanity.com/paper/2205.05124/) (2022), cited by 32 papers here - [Personalized Steering of Large Language Models: Versatile Steering Vectors Through Bi-directional Preference Optimization](https://mlsanity.com/paper/2406.00045/) (2024), cited by 29 papers here - [Improving Instruction-Following in Language Models through Activation Steering](https://mlsanity.com/paper/2410.12877/) (2024), cited by 28 papers here - [Semantics-Adaptive Activation Intervention for LLMs via Dynamic Steering Vectors](https://mlsanity.com/paper/2410.12299/) (2024), cited by 21 papers here - [A Unified Understanding and Evaluation of Steering Methods](https://mlsanity.com/paper/2502.02716/) (2025), cited by 19 papers here - [Understanding (Un)Reliability of Steering Vectors in Language Models](https://mlsanity.com/paper/2505.22637/) (2025), cited by 17 papers here - [Extending Activation Steering to Broad Skills and Multiple Behaviours](https://mlsanity.com/paper/2403.05767/) (2024), cited by 15 papers here ## New (24 Sep–14 Oct) - [Causal Syntax Circuits in CodeLMs: Budget and Probe Decide What Looks Localised](https://mlsanity.com/paper/2610.20683/) (2026-10-14) - [Genesis Steering: Better Activation Steering Through Training Data](https://mlsanity.com/paper/2610.19114/) (2026-10-13) - [Perturbation, Not Provenance: Why Activation Steering Does Not Make a Robust Text Watermark](https://mlsanity.com/paper/2610.17810/) (2026-10-13) - [Beyond Bias Scores: Answerability Confounds in Activation Steering with Gradient-Constrained Direction Design (GCD)](https://mlsanity.com/paper/2610.17641/) (2026-10-13) - [The Sharp Directions Are Against You: Curvature Analysis of Activation Steering](https://mlsanity.com/paper/2610.16084/) (2026-10-12) - [What Survives When Activation Steering Is Used as Memory](https://mlsanity.com/paper/2610.14958/) (2026-10-10) - [Steerspeech: Activation Steering For Emotion Control In Generated Speech](https://mlsanity.com/paper/2610.10415/) (2026-10-07) - [Steering Follows Geometry, Not Labels: Emotion Directions in a Full-Duplex Speech Model](https://mlsanity.com/paper/2610.08887/) (2026-10-06) - [Does Steering Break Your Model? A Multi-Dimensional Evaluation Suite for LLM Steering Methods](https://mlsanity.com/paper/2610.07722/) (2026-10-06) - [Loud and Clear: Dynamic Activation Steering for Improving Speech Intelligibility in Noisy Environments](https://mlsanity.com/paper/2610.07647/) (2026-10-06) - [Steering by Influence: Curvature Aware Data Weighting for Activation Steering](https://mlsanity.com/paper/2610.06383/) (2026-10-05) - [Tracing a Sparse Emotion-Control Circuit in LLM-Based Text-to-Speech](https://mlsanity.com/paper/2610.05080/) (2026-10-04) - [One Token Can Be Enough: Bridging Prompting and Activation Steering with Prefix Steering](https://mlsanity.com/paper/2610.04967/) (2026-10-04) - [Why Subliminal Learning Needs So Much Data: A Noisy Inverse View through Steering Vector Recovery](https://mlsanity.com/paper/2610.04907/) (2026-10-04) - [Steering Speech-Language Models: Training-Free Task Specialization via Contrastive Activation Addition](https://mlsanity.com/paper/2610.04683/) (2026-10-03) - [First-Order Steering: Translating Weight Adaptation into Activation Steering](https://mlsanity.com/paper/2610.04283/) (2026-10-03) - [On the Steering Dimensionality of Refusal in Language Models](https://mlsanity.com/paper/2610.04245/) (2026-10-03) - [Predicting Steering Vectors and Adapter Weights for Few-Shot Author-Style Transfer](https://mlsanity.com/paper/2610.03163/) (2026-10-02) - [Hesitation Has a Geometry: Entropy-Trained Hyperbolic Probes for Sparse Activation Steering](https://mlsanity.com/paper/2610.02391/) (2026-10-01) - [Kernelized Activation Steering](https://mlsanity.com/paper/2610.01062/) (2026-10-01) - [Component and Dimension Sparsity in Transformer Refusal Mechanisms](https://mlsanity.com/paper/2610.06903/) (2026-09-30) - [Steering Fields: Adaptive Vector Fields for Safe Image Generation and Beyond](https://mlsanity.com/paper/2609.39573/) (2026-09-30) - [Contrastive Covariance Subspaces for Activation Steering](https://mlsanity.com/paper/2610.14025/) (2026-09-29) - [Persona Dosing: Calibrated Activation Steering for Graded Trait Control](https://mlsanity.com/paper/2609.36388/) (2026-09-28) - [Controlling Speaking Rate in Autoregressive TTS via Activation Steering](https://mlsanity.com/paper/2609.33810/) (2026-09-27) - [LocUS: Head Selection and Subspace Projection for Targeted Activation Steering](https://mlsanity.com/paper/2609.31122/) (2026-09-25) - [The Linear Representation Hypothesis for Vision-Language-Action Models](https://mlsanity.com/paper/2609.30996/) (2026-09-25) - [Adaptive Multi-Value Control in LLMs via Causal Activation Steering](https://mlsanity.com/paper/2609.30405/) (2026-09-24)