LoRA-as-an-Attack! Piercing LLM Safety Under The Share-and-Play Scenario

arXiv:2403.00108v1 Announce Type: cross
Abstract: Fine-tuning LLMs is crucial to enhancing their task-specific performance and ensuring model behaviors are aligned with human preferences. Among various fine-tuning methods, LoRA is popular for its efficiency and ease to use, allowing end-users to easily post and adopt lightweight LoRA modules on open-source platforms to tailor their model for different customization. However, such a handy share-and-play setting opens up new attack surfaces, that the attacker can render LoRA as an attacker, such as backdoor …

abstract arxiv cs.ai cs.cl cs.cr efficiency fine-tuning human llm llms lora modules performance platforms popular safety type