9 papers
Graphs Don't Stay Secret: Practical Subgraph Reconstruction Attacks on Defended Graph RAG
Minkyoo Song, Jaehan Kim, Myungchul Kang +3
Graph-based retrieval-augmented generation (Graph RAG) is increasingly deployed to support LLM applications by augmenting user queries with structured knowledge retrieved from a kn…
SkillMutator: Benchmarking and Defending Language-and-Code Cross-modal Attacks on LLM Agent Skills
Youngduk Kim, Minkyoo Song, Seungwon Shin
Large language model (LLM) agents increasingly extend their capabilities at runtime by loading Agent Skills, which pair natural-language specifications (SKILL.md) with executable s…
: Politically Controversial Content Generation via Jailbreaking Attacks on GPT-based Text-to-Image Models
Wonwoo Choi, Minjae Seo, Minkyoo Song +3
The rapid evolution of text-to-image (T2I) models has enabled high-fidelity visual synthesis on a global scale. However, these advancements have introduced significant security ris…
PassREfinder-FL: Privacy-Preserving Credential Stuffing Risk Prediction via Graph-Based Federated Learning for Representing Password Reuse between Websites
Jaehan Kim, Minkyoo Song, Minjae Seo +3
Credential stuffing attacks have caused significant harm to online users who frequently reuse passwords across multiple websites. While prior research has attempted to detect users…
Defending MoE LLMs against Harmful Fine-Tuning via Safety Routing Alignment
Jaehan Kim, Minkyoo Song, Seungwon Shin +1
Recent large language models (LLMs) have increasingly adopted the Mixture-of-Experts (MoE) architecture for efficiency. MoE-based LLMs heavily depend on a superficial safety mechan…
Covering Cracks in Content Moderation: Delexicalized Distant Supervision for Illicit Drug Jargon Detection
Minkyoo Song, Eugene Jang, Jaehan Kim +1
In light of rising drug-related concerns and the increasing role of social media, sales and discussions of illicit drugs have become commonplace online. Social media platforms host…