Multilingual Safety Alignment via Self-Distillation

arXiv:2605.02971 · cs.LG, cs.AI, cs.CL · Submitted 2026-05-03 · Read on arXiv

cs.LG, cs.AI, cs.CL

Submitted: 2026-05-03

Updated: 2026-10-18

Comments: Accepted by NeurIPS 2026

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Distillation — 4× in this paper · explained in 2 episodes
• Self-Distillation — 3× in this paper · explained in 7 episodes
• Benchmark — 1× in this paper · explained in 2 episodes
• Dual-Perspective Safety Weighting — 1× in this paper
• Large Language Models (LLMs) — 1× in this paper · explained in 8 episodes
• Low-Resource Languages — 1× in this paper · explained in 4 episodes
• Multilingual Self-Distillation — 1× in this paper

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