Cover Image

Here's a Free Lunch: Sanitizing Backdoored Models with Model Merge

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Arora, Ansh, He, Xuanli, Mozes, Maximilian, Swain, Srinibas, Dras, Mark, Xu, Qiongkai
Format:	Preprint
Published:	2024
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Defending against Backdoor Attacks via Module Switching
by: Li, Weijun, et al.
Published: (2025)

Backdoor Attack on Multilingual Machine Translation
by: Wang, Jun, et al.
Published: (2024)

Cut the Deadwood Out: Backdoor Purification via Guided Module Substitution
by: Tong, Yao, et al.
Published: (2024)

SEEP: Training Dynamics Grounds Latent Representation Search for Mitigating Backdoor Poisoning Attacks
by: He, Xuanli, et al.
Published: (2024)

Seeing the Forest through the Trees: Data Leakage from Partial Transformer Gradients
by: Li, Weijun, et al.
Published: (2024)

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

Attacks on Third-Party APIs of Large Language Models
by: Zhao, Wanru, et al.
Published: (2024)

Beyond Theoretical Bounds: Empirical Privacy Loss Calibration for Text Rewriting Under Local Differential Privacy
by: Li, Weijun, et al.
Published: (2026)

WARDEN: Multi-Directional Backdoor Watermarks for Embedding-as-a-Service Copyright Protection
by: Shetty, Anudeex, et al.
Published: (2024)

IDT: Dual-Task Adversarial Attacks for Privacy Protection
by: Faustini, Pedro, et al.
Published: (2024)

Seeing the Threat: Vulnerabilities in Vision-Language Models to Adversarial Attack
by: Ren, Juan, et al.
Published: (2025)

AlignCultura: Towards Culturally Aligned Large Language Models?
by: Kashyap, Gautam Siddharth, et al.
Published: (2026)

When the Model Said 'No Comment', We Knew Helpfulness Was Dead, Honesty Was Alive, and Safety Was Terrified
by: Kashyap, Gautam Siddharth, et al.
Published: (2026)

CogMem: A Cognitive Memory Architecture for Sustained Multi-Turn Reasoning in Large Language Models
by: Zhang, Yiran, et al.
Published: (2025)

ACE-Merging: Data-Free Model Merging with Adaptive Covariance Estimation
by: Xu, Bo, et al.
Published: (2026)

Myanmar XNLI: Building a Dataset and Exploring Low-resource Approaches to Natural Language Inference with Myanmar
by: Htet, Aung Kyaw, et al.
Published: (2025)

Generative Models are Self-Watermarked: Declaring Model Authentication through Re-Generation
by: Desu, Aditya, et al.
Published: (2024)

SHIELD: Classifier-Guided Prompting for Robust and Safer LVLMs
by: Ren, Juan, et al.
Published: (2025)

Should LLM Safety Be More Than Refusing Harmful Instructions?
by: Maskey, Utsav, et al.
Published: (2025)

Steering Over-refusals Towards Safety in Retrieval Augmented Generation
by: Maskey, Utsav, et al.
Published: (2025)

Over-Refusal and Representation Subspaces: A Mechanistic Analysis of Task-Conditioned Refusal in Aligned LLMs
by: Maskey, Utsav, et al.
Published: (2026)

Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch
by: Yu, Le, et al.
Published: (2023)

Leftover Lunch: Advantage-based Offline Reinforcement Learning for Language Models
by: Baheti, Ashutosh, et al.
Published: (2023)

Knowledge Sanitization of Large Language Models
by: Ishibashi, Yoichi, et al.
Published: (2023)

Adaptive Cost-Efficient Evaluation for Reliable Patent Claim Generation
by: Yoo, Yongmin, et al.
Published: (2026)

We Think, Therefore We Align LLMs to Helpful, Harmless and Honest Before They Go Wrong
by: Kashyap, Gautam Siddharth, et al.
Published: (2025)

Too Helpful, Too Harmless, Too Honest or Just Right?
by: Kashyap, Gautam Siddharth, et al.
Published: (2025)

The Empirical Impact of Data Sanitization on Language Models
by: Pal, Anwesan, et al.
Published: (2024)

Is It a Free Lunch for Removing Outliers during Pretraining?
by: Liao, Baohao, et al.
Published: (2024)

Steering Towards Fairness: Mitigating Political Bias in LLMs
by: Nadeem, Afrozah, et al.
Published: (2025)

Fairness Evaluation and Inference Level Mitigation in LLMs
by: Nadeem, Afrozah, et al.
Published: (2025)

Framing Political Bias in Multilingual LLMs Across Pakistani Languages
by: Nadeem, Afrozah, et al.
Published: (2025)

Free Lunch for Pass@$k$? Low Cost Diverse Sampling for Diffusion Language Models
by: Lamont, Sean, et al.
Published: (2026)

Merging Triggers, Breaking Backdoors: Defensive Poisoning for Instruction-Tuned Language Models
by: Kim, San, et al.
Published: (2026)

Beyond the Black Box: Demystifying Multi-Turn LLM Reasoning with VISTA
by: Zhang, Yiran, et al.
Published: (2025)

SafeConstellations: Mitigating Over-Refusals in LLMs Through Task-Aware Representation Steering
by: Maskey, Utsav, et al.
Published: (2025)

Is Long-to-Short a Free Lunch? Investigating Inconsistency and Reasoning Efficiency in LRMs
by: Yang, Shu, et al.
Published: (2025)

PatentScore: Multi-dimensional Evaluation of LLM-Generated Patent Claims
by: Yoo, Yongmin, et al.
Published: (2025)

Asking LLMs to Verify First is Almost Free Lunch
by: Wu, Shiguang, et al.
Published: (2025)

LLMInit: A Free Lunch from Large Language Models for Selective Initialization of Recommendation
by: Zhang, Weizhi, et al.
Published: (2025)