RogueMerge is a unified attack method that jointly optimizes task vectors to succeed after merging, using stochastic min-max simulation for unknown merging settings and a Taylor-approximated DRO for prompt generalization on generative LLMs.
Backdoorllm: A comprehensive benchmark for backdoor attacks on large language models
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
SHADOWMASK backdoors MDLMs by replacing the all-mask terminal distribution with a trigger-mask mixture prior, achieving near-100% attack success on DiT and LLaDA-8B models across multiple datasets while resisting fine-tuning and some defenses.
SkillTrojan backdoors skill-based agents by partitioning an encrypted payload across benign-looking skills that reassemble and execute only under a predefined trigger.
Privacy attacks on LLMs show strong signals for membership inference and backdoors but weaker performance for attribute inference and data extraction, with risks highly dependent on system configuration.
A framework detects LLM anomalies including hallucinations, jailbreaks, and backdoors by forensic inspection of layer-wise hidden state patterns, reporting over 95% accuracy with minimal computational overhead.
Survey of harmful fine-tuning attacks on LLMs, their variants, defense strategies, mechanical analysis, and evaluation methodologies.
citing papers explorer
-
RogueMerge: Robust and Unified Attacks against LLM Model Merging
RogueMerge is a unified attack method that jointly optimizes task vectors to succeed after merging, using stochastic min-max simulation for unknown merging settings and a Taylor-approximated DRO for prompt generalization on generative LLMs.
-
Backdooring Masked Diffusion Language Models
SHADOWMASK backdoors MDLMs by replacing the all-mask terminal distribution with a trigger-mask mixture prior, achieving near-100% attack success on DiT and LLaDA-8B models across multiple datasets while resisting fine-tuning and some defenses.
-
SkillTrojan: Backdoor Attacks on Skill-Based Agent Systems
SkillTrojan backdoors skill-based agents by partitioning an encrypted payload across benign-looking skills that reassemble and execute only under a predefined trigger.
-
On the Privacy of LLMs: An Ablation Study
Privacy attacks on LLMs show strong signals for membership inference and backdoors but weaker performance for attribute inference and data extraction, with risks highly dependent on system configuration.
-
Exposing the Ghost in the Transformer: Abnormal Detection for Large Language Models via Hidden State Forensics
A framework detects LLM anomalies including hallucinations, jailbreaks, and backdoors by forensic inspection of layer-wise hidden state patterns, reporting over 95% accuracy with minimal computational overhead.
-
Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey
Survey of harmful fine-tuning attacks on LLMs, their variants, defense strategies, mechanical analysis, and evaluation methodologies.