Cover Image

Safety Tax: Safety Alignment Makes Your Large Reasoning Models Less Reasonable

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Huang, Tiansheng, Hu, Sihao, Ilhan, Fatih, Tekin, Selim Furkan, Yahn, Zachary, Xu, Yichang, Liu, Ling
Format:	Preprint
Published:	2025
Subjects:	Cryptography and Security Artificial Intelligence Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey
by: Huang, Tiansheng, et al.
Published: (2024)

Virus: Harmful Fine-tuning Attack for Large Language Models Bypassing Guardrail Moderation
by: Huang, Tiansheng, et al.
Published: (2025)

MELT: A Behavioral Trace Dataset for High-Risk Memecoin Launch Detection
by: Hu, Sihao, et al.
Published: (2026)

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

Large Language Model based Smart Contract Auditing with LLMBugScanner
by: Yuan, Yining, et al.
Published: (2025)

H3Fusion: Helpful, Harmless, Honest Fusion of Aligned LLMs
by: Tekin, Selim Furkan, et al.
Published: (2024)

Adversarial Attention Perturbations for Large Object Detection Transformers
by: Yahn, Zachary, et al.
Published: (2025)

A Multi-Agent Perception-Action Alliance for Efficient Long Video Reasoning
by: Xu, Yichang, et al.
Published: (2026)

A Survey on Large Language Model-Based Game Agents
by: Hu, Sihao, et al.
Published: (2024)

A Neurosymbolic Agent System for Compositional Visual Reasoning
by: Xu, Yichang, et al.
Published: (2025)

Personalized Face Privacy Protection From a Single Image
by: Yahn, Zachary, et al.
Published: (2026)

Antidote: Post-fine-tuning Safety Alignment for Large Language Models against Harmful Fine-tuning
by: Huang, Tiansheng, et al.
Published: (2024)

Booster: Tackling Harmful Fine-tuning for Large Language Models via Attenuating Harmful Perturbation
by: Huang, Tiansheng, et al.
Published: (2024)

Attention-aware Inference Optimizations for Large Vision-Language Models with Memory-efficient Decoding
by: Ilhan, Fatih, et al.
Published: (2026)

Colluding LoRA: A Compositional Vulnerability in LLM Safety Alignment
by: Ding, Sihao
Published: (2026)

LLM-TOPLA: Efficient LLM Ensemble by Maximising Diversity
by: Tekin, Selim Furkan, et al.
Published: (2024)

AutoRAN: Automated Hijacking of Safety Reasoning in Large Reasoning Models
by: Liang, Jiacheng, et al.
Published: (2025)

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

Reasoning Hijacking: The Fragility of Reasoning Alignment in Large Language Models
by: Liu, Yuansen, et al.
Published: (2026)

Self-Jailbreaking: Language Models Can Reason Themselves Out of Safety Alignment After Benign Reasoning Training
by: Yong, Zheng-Xin, et al.
Published: (2025)

Refusal Falls off a Cliff: How Safety Alignment Fails in Reasoning?
by: Yin, Qingyu, et al.
Published: (2025)

SafeThinker: Reasoning about Risk to Deepen Safety Beyond Shallow Alignment
by: Fang, Xianya, et al.
Published: (2026)

PRISM: Robust VLM Alignment with Principled Reasoning for Integrated Safety in Multimodality
by: Li, Nanxi, et al.
Published: (2025)

GSPR: Aligning LLM Safeguards as Generalizable Safety Policy Reasoners
by: Li, Haoran, et al.
Published: (2025)

Beyond Content Safety: Real-Time Monitoring for Reasoning Vulnerabilities in Large Language Models
by: Wang, Xunguang, et al.
Published: (2026)

RAJ-PGA: Reasoning-Activated Jailbreak and Principle-Guided Alignment Framework for Large Reasoning Models
by: Chen, Jianhao, et al.
Published: (2025)

Safety Context Injection: Inference-Time Safety Alignment via Static Filtering and Agentic Analysis
by: Xu, Zhenhao, et al.
Published: (2026)

Conflicts Make Large Reasoning Models Vulnerable to Attacks
by: Liu, Honghao, et al.
Published: (2026)

Enhancing Model Defense Against Jailbreaks with Proactive Safety Reasoning
by: Yang, Xianglin, et al.
Published: (2025)

SafeMLRM: Demystifying Safety in Multi-modal Large Reasoning Models
by: Fang, Junfeng, et al.
Published: (2025)

Structured Visual Narratives Undermine Safety Alignment in Multimodal Large Language Models
by: Tan, Rui Yang, et al.
Published: (2026)

The Verifier Tax: Horizon Dependent Safety Success Tradeoffs in Tool Using LLM Agents
by: Sah, Tanmay, et al.
Published: (2026)

Safety is Not Only About Refusal: Reasoning-Enhanced Fine-tuning for Interpretable LLM Safety
by: Zhang, Yuyou, et al.
Published: (2025)

EnchTable: Unified Safety Alignment Transfer in Fine-tuned Large Language Models
by: Wu, Jialin, et al.
Published: (2025)

ReasAlign: Reasoning Enhanced Safety Alignment against Prompt Injection Attack
by: Li, Hao, et al.
Published: (2026)

Adversary-Aware DPO: Enhancing Safety Alignment in Vision Language Models via Adversarial Training
by: Weng, Fenghua, et al.
Published: (2025)

Bag of Tricks for Subverting Reasoning-based Safety Guardrails
by: Chen, Shuo, et al.
Published: (2025)

What Matters For Safety Alignment?
by: Li, Xing, et al.
Published: (2026)

Be Your Own Red Teamer: Safety Alignment via Self-Play and Reflective Experience Replay
by: Wang, Hao, et al.
Published: (2026)

When Reasoning Leaks Membership: Membership Inference Attack on Black-box Large Reasoning Models
by: Hu, Ruihan, et al.
Published: (2026)