Activation Steering for Aligned Open-ended Generation without Sacrificing Coherence
Alignment in LLMs is more brittle than commonly assumed: misalignment can be induced by adversarial prompts, benign fine-tuning, emergent misalignment, and goal misgeneralization. Recent evidence suggests that some misalignment behaviors are encoded as linear structure in activation space, making it tractable via activation steering, which could be used as a lightweight runtime defense. We implement three methods: Steer-With-Fixed-Coefficient (SwFC), which applies uniform additive steering, and
Record details
Published: 9 April 2026
Source: arXiv
Category: Research
Topics: Safety & alignment · Military & security
Retrieved: 14 July 2026
Related evidence
These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.
AI Integrity: A New Paradigm for Verifiable AI Governance
arXiv · 13 April 2026
GUARD-SLM: Token Activation-Based Defense Against Jailbreak Attacks for Small Language Models
arXiv · 28 March 2026
The Autonomy Tax: Defense Training Breaks LLM Agents
arXiv · 19 March 2026
SFCoT: Safer Chain-of-Thought via Active Safety Evaluation and Calibration
arXiv · 16 March 2026
Directional Embedding Smoothing for Robust Vision Language Models
arXiv · 16 March 2026
Disentangling Intent from Role: Adversarial Self-Play for Persona-Invariant Safety Alignment
arXiv · 3 May 2026
How to cite this record
ethics.ai (9 April 2026), “Activation Steering for Aligned Open-ended Generation without Sacrificing Coherence,” evidence record 6131, https://ethics.ai/record/6131 (originally published by arXiv).
Use and limitations
This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.