SOURCE-LINKED INTELLIGENCE

Does Fine-Tuning Undo Activation Steering? Behavioural Recovery Without Weight-Edit Reversal

arXiv · AI, language, vision and robotics · article · Aug 25, 2026 · UTC

Activation steering can be embedded directly into a language model's weights, shaping behaviour without inference-time intervention and offering a way to encode alignment prior to release. However, models are routinely fine-tuned after deployment, and it is unknown whether embedded interventions survive this. We study the stability of embedded steering for refusal suppression and brevity induction across five instruction-tuned models (3B-14B) under non-adversarial SFT and RLHF. Behaviourally, preservation tracks the training data: steering degrades when optimisation pressure contradicts the ta

Read original source ↗ Open in workspace

recordType

paper

region

Global