Jun Sun
4 papers · Latest:
Layerwise Convergence Fingerprints for Runtime Misbehavior Detection in Large Language Models
LCF is a tuning-free runtime monitor that detects LLM misbehavior like backdoors, jailbreaks, and prompt injections by analyzing hidden-state trajectories.
Train in Vain: Functionality-Preserving Poisoning to Prevent Unauthorized Use of Code Datasets
FunPoison introduces a functionality-preserving poisoning method to prevent unauthorized use of code datasets for training CodeLLMs, maintaining compilability.
ClawGuard: A Runtime Security Framework for Tool-Augmented LLM Agents Against Indirect Prompt Injection
ClawGuard is a runtime security framework protecting tool-augmented LLM agents from indirect prompt injection by enforcing rules at tool-call boundaries.
The Salami Slicing Threat: Exploiting Cumulative Risks in LLM Systems
This paper introduces Salami Slicing, a novel multi-turn jailbreak attack that exploits cumulative low-risk inputs to bypass LLM safety, achieving high success rates.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.