Cover Image

Safety Alignment Depth in Large Language Models: A Markov Chain Perspective

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Kao, Ching-Chia, Yu, Chia-Mu, Lu, Chun-Shien, Chen, Chu-Song
Format:	Preprint
Published:	2025
Subjects:	Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

The Great Contradiction Showdown: How Jailbreak and Stealth Wrestle in Vision-Language Models?
by: Kao, Ching-Chia, et al.
Published: (2024)

Safe LoRA: the Silver Lining of Reducing Safety Risks when Fine-tuning Large Language Models
by: Hsu, Chia-Yi, et al.
Published: (2024)

Defending Against Repetitive Backdoor Attacks on Semi-supervised Learning through Lens of Rate-Distortion-Perception Trade-off
by: Lee, Cheng-Yi, et al.
Published: (2024)

Harmless Yet Harmful: Neutral Prompting Attacks for Stealthy Hallucination Steering in Agent Skills
by: Hsu, Chia-Yi, et al.
Published: (2026)

Counter-Current Learning: A Biologically Plausible Dual Network Approach for Deep Learning
by: Kao, Chia-Hsiang, et al.
Published: (2024)

Feature Alignment and Representation Transfer in Knowledge Distillation for Large Language Models
by: Yang, Junjie, et al.
Published: (2025)

Ring-A-Bell! How Reliable are Concept Removal Methods for Diffusion Models?
by: Tsai, Yu-Lin, et al.
Published: (2023)

Open-Vocabulary Panoptic Segmentation Using BERT Pre-Training of Vision-Language Multiway Transformer Model
by: Chen, Yi-Chia, et al.
Published: (2024)

Large Language Models as Markov Chains
by: Zekri, Oussama, et al.
Published: (2024)

Test-time Adversarial Defense with Opposite Adversarial Path and High Attack Time Cost
by: Yeh, Cheng-Han, et al.
Published: (2024)

A Survey of Research in Large Language Models for Electronic Design Automation
by: Pan, Jingyu, et al.
Published: (2025)

The Curse of Depth in Large Language Models
by: Sun, Wenfang, et al.
Published: (2025)

Alignment and Safety in Large Language Models: Safety Mechanisms, Training Paradigms, and Emerging Challenges
by: Lu, Haoran, et al.
Published: (2025)

OCEAN: Offline Chain-of-thought Evaluation and Alignment in Large Language Models
by: Wu, Junda, et al.
Published: (2024)

Minimal Cascade Gradient Smoothing for Fast Transferable Preemptive Adversarial Defense
by: Wang, Hanrui, et al.
Published: (2024)

GreedyPixel: Fine-Grained Black-Box Adversarial Attack Via Greedy Algorithm
by: Wang, Hanrui, et al.
Published: (2025)

Targeted Vaccine: Safety Alignment for Large Language Models against Harmful Fine-Tuning via Layer-wise Perturbation
by: Liu, Guozhi, et al.
Published: (2024)

Bridging Compressed Image Latents and Multimodal Large Language Models
by: Kao, Chia-Hao, et al.
Published: (2024)

Any-Depth Alignment: Unlocking Innate Safety Alignment of LLMs to Any-Depth
by: Zhang, Jiawei, et al.
Published: (2025)

Model Reprogramming Demystified: A Neural Tangent Kernel Perspective
by: Chung, Ming-Yu, et al.
Published: (2025)

VP-NTK: Exploring the Benefits of Visual Prompting in Differentially Private Data Synthesis
by: Hsu, Chia-Yi, et al.
Published: (2025)

Controlling Exploration-Exploitation in GFlowNets via Markov Chain Perspectives
by: Chen, Lin, et al.
Published: (2026)

SafeNeuron: Neuron-Level Safety Alignment for Large Language Models
by: Wang, Zhaoxin, et al.
Published: (2026)

Tight Analysis of Decentralized SGD: A Markov Chain Perspective
by: Versini, Lucas, et al.
Published: (2026)

Pharmacist: Safety Alignment Data Curation for Large Language Models against Harmful Fine-tuning
by: Liu, Guozhi, et al.
Published: (2025)

A Probabilistic Perspective on Unlearning and Alignment for Large Language Models
by: Scholten, Yan, et al.
Published: (2024)

Rethinking Backdoor Attacks on Dataset Distillation: A Kernel Method Perspective
by: Chung, Ming-Yu, et al.
Published: (2023)

BADTV: Unveiling Backdoor Threats in Third-Party Task Vectors
by: Hsu, Chia-Yi, et al.
Published: (2025)

Emulated Disalignment: Safety Alignment for Large Language Models May Backfire!
by: Zhou, Zhanhui, et al.
Published: (2024)

Rethinking Large Language Model Distillation: A Constrained Markov Decision Process Perspective
by: Zimmer, Matthieu, et al.
Published: (2025)

Layer-Aware Task Arithmetic: Disentangling Task-Specific and Instruction-Following Knowledge
by: Chen, Yan-Lun, et al.
Published: (2025)

Caduceus: Bi-Directional Equivariant Long-Range DNA Sequence Modeling
by: Schiff, Yair, et al.
Published: (2024)

Conformal Tail Risk Control for Large Language Model Alignment
by: Chen, Catherine Yu-Chi, et al.
Published: (2025)

DiffMI: Breaking Face Recognition Privacy via Diffusion-Driven Training-Free Model Inversion
by: Wang, Hanrui, et al.
Published: (2025)

Q-resafe: Assessing Safety Risks and Quantization-aware Safety Patching for Quantized Large Language Models
by: Chen, Kejia, et al.
Published: (2025)

A Modularized Framework for Piecewise-Stationary Restless Bandits
by: Li, Kuan-Ta, et al.
Published: (2026)

QG-CoC: Question-Guided Chain-of-Captions for Large Multimodal Models
by: Kao, Kuei-Chun, et al.
Published: (2025)

Lisa: Lazy Safety Alignment for Large Language Models against Harmful Fine-tuning Attack
by: Huang, Tiansheng, et al.
Published: (2024)

Progressive Alignment with VLM-LLM Feature to Augment Defect Classification for the ASE Dataset
by: Hsu, Chih-Chung, et al.
Published: (2024)

Solving for X and Beyond: Can Large Language Models Solve Complex Math Problems with More-Than-Two Unknowns?
by: Kao, Kuei-Chun, et al.
Published: (2024)