Cover Image

DataShield: Safety-degrading Data Filtering for LLM Benign Instruction Fine-Tuning

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Zhang, Junbo, Zhou, Qianli, Deng, Xinyang, Jiang, Wen, Pan, Jie, Zhu, Jinbiao
Format:	Preprint
Published:	2026
Subjects:	Cryptography and Security Artificial Intelligence Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Understanding and Mitigating Over-refusal for Large Language Models via Safety Representation
by: Zhang, Junbo, et al.
Published: (2025)

Cross-Task Defense: Instruction-Tuning LLMs for Content Safety
by: Fu, Yu, et al.
Published: (2024)

Sugar-Coated Poison: Benign Generation Unlocks LLM Jailbreaking
by: Wu, Yu-Hang, et al.
Published: (2025)

What is in Your Safe Data? Identifying Benign Data that Breaks Safety
by: He, Luxi, et al.
Published: (2024)

Benign Fine-Tuning Breaks Safety Alignment in Audio LLMs
by: Roh, Jaechul, et al.
Published: (2026)

Generalization-Enhanced Code Vulnerability Detection via Multi-Task Instruction Fine-Tuning
by: Du, Xiaohu, et al.
Published: (2024)

SCOUT: A Defense Against Data Poisoning Attacks in Fine-Tuned Language Models
by: Afane, Mohamed, et al.
Published: (2025)

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

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

Instructional Segment Embedding: Improving LLM Safety with Instruction Hierarchy
by: Wu, Tong, et al.
Published: (2024)

OneShield -- the Next Generation of LLM Guardrails
by: DeLuca, Chad, et al.
Published: (2025)

One Word at a Time: Incremental Completion Decomposition Breaks LLM Safety
by: Arif, Samee, et al.
Published: (2026)

Token-level Data Selection for Safe LLM Fine-tuning
by: Li, Yanping, et al.
Published: (2026)

Secure LLM Fine-Tuning via Safety-Aware Probing
by: Wu, Chengcan, et al.
Published: (2025)

AgentShield: Deception-based Compromise Detection for Tool-using LLM Agents
by: Rassul, Yassin H., et al.
Published: (2026)

Defending Against Weight-Poisoning Backdoor Attacks for Parameter-Efficient Fine-Tuning
by: Zhao, Shuai, et al.
Published: (2024)

Layer-Aware Representation Filtering: Purifying Finetuning Data to Preserve LLM Safety Alignment
by: Li, Hao, et al.
Published: (2025)

AdaSteer: Your Aligned LLM is Inherently an Adaptive Jailbreak Defender
by: Zhao, Weixiang, et al.
Published: (2025)

The Art of (Mis)alignment: How Fine-Tuning Methods Effectively Misalign and Realign LLMs in Post-Training
by: Zhang, Rui, et al.
Published: (2026)

TrojanPraise: Jailbreak LLMs via Benign Fine-Tuning
by: Xie, Zhixin, et al.
Published: (2026)

TuBA: Cross-Lingual Transferability of Backdoor Attacks in LLMs with Instruction Tuning
by: He, Xuanli, et al.
Published: (2024)

Rethinking Bottlenecks in Safety Fine-Tuning of Vision Language Models
by: Ding, Yi, et al.
Published: (2025)

A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deployment
by: Wang, Kun, et al.
Published: (2025)

Leaner Training, Lower Leakage: Revisiting Memorization in LLM Fine-Tuning with LoRA
by: Wang, Fei, et al.
Published: (2025)

Privacy-Preserving Parameter-Efficient Fine-Tuning for Large Language Model Services
by: Li, Yansong, et al.
Published: (2023)

Behavioral Canaries: Auditing Private Retrieved Context Usage in RL Fine-Tuning
by: Chen, Chaoran, et al.
Published: (2026)

Mitigating Fine-tuning based Jailbreak Attack with Backdoor Enhanced Safety Alignment
by: Wang, Jiongxiao, et al.
Published: (2024)

Why LLM Safety Guardrails Collapse After Fine-tuning: A Similarity Analysis Between Alignment and Fine-tuning Datasets
by: Hsiung, Lei, et al.
Published: (2025)

RewardDS: Privacy-Preserving Fine-Tuning for Large Language Models via Reward Driven Data Synthesis
by: Wang, Jianwei, et al.
Published: (2025)

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

Emerging Safety Attack and Defense in Federated Instruction Tuning of Large Language Models
by: Ye, Rui, et al.
Published: (2024)

SafeTune: Mitigating Data Poisoning in LLM Fine-Tuning for RTL Code Generation
by: Rezakhani, Mahshid, et al.
Published: (2026)

Root Defence Strategies: Ensuring Safety of LLM at the Decoding Level
by: Zeng, Xinyi, et al.
Published: (2024)

CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning
by: Yi, Biao, et al.
Published: (2025)

PandaGuard: Systematic Evaluation of LLM Safety against Jailbreaking Attacks
by: Shen, Guobin, et al.
Published: (2025)

DAVE: A Policy-Enforcing LLM Spokesperson for Secure Multi-Document Data Sharing
by: Brinkhege, René, et al.
Published: (2026)

Analysing Safety Risks in LLMs Fine-Tuned with Pseudo-Malicious Cyber Security Data
by: ElZemity, Adel, et al.
Published: (2025)

Safety Anchor: Defending Harmful Fine-tuning via Geometric Bottlenecks
by: Lu, Guoxin, et al.
Published: (2026)

Refining Input Guardrails: Enhancing LLM-as-a-Judge Efficiency Through Chain-of-Thought Fine-Tuning and Alignment
by: Rad, Melissa Kazemi, et al.
Published: (2025)

TWGuard: A Case Study of LLM Safety Guardrails for Localized Linguistic Contexts
by: Chu, Hua-Rong, et al.
Published: (2026)