D-STEER: Two-Way Control of DPO-Induced Safety Behavior in Activation Space

arXiv:2512.11838 · cs.LG · Submitted 2025-12-03 · Read on arXiv

cs.LG

Submitted: 2025-12-03

Updated: 2026-10-16

Terminology

• Steering — 2× in this paper · explained in 2 episodes
• Direct Preference Optimization — 1× in this paper
• Direct Preference Optimization (DPO) — 1× in this paper · explained in 10 episodes
• Embeddings — 1× in this paper · explained in 2 episodes
• Entropy Collapse — 1× in this paper · explained in 2 episodes

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