Kernelized Activation Steering

arXiv:2610.01062 · cs.LG · Submitted 2026-10-01 · Read on arXiv

cs.LG

Submitted: 2026-10-01

Updated: 2026-10-03

Code: https://github.com/black-forest-labs/flux

Terminology

• Activation Steering — 1× in this paper · explained in 4 episodes
• Steering — 1× in this paper · explained in 2 episodes

Sources

• Understanding intermediate layers using linear classifier probes
• Microsoft COCO Captions: Data Collection and Evaluation Server
• Fundamental Limits of Perfect Concept Erasure
• The Pile: An 800GB Dataset of Diverse Text for Language Modeling
• Curveball Steering: The Right Direction To Steer Isn't Always Linear
• Nonlinear Concept Erasure: a Density Matching Approach
• Steering Language Models With Activation Engineering
• Representation Engineering: A Top-Down Approach to AI Transparency

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks