Tamper-Resistant Safeguards for Open-Weight LLMs
CAIS · University of California, San Diego · Center for AI Safety · Intology · Department of Computer Science · Carnegie Mellon University · Harvard University · Intology AI · CMU, Carnegie Mellon University · Berkeley · University of Illinois, Urbana Champaign · UC Berkeley · University of Illinois, Urbana-Champaign
PDF 由论文原始站点提供，PaperCompass 不保存论文文件。
摘要
Rapid advances in the capabilities of large language models (LLMs) have raised widespread concerns regarding their potential for malicious use. Open-weight LLMs present unique challenges, as existing safeguards lack robustness to tampering attacks that modify model weights. For example, recent works have demonstrated that refusal and unlearning safeguards can be trivially removed with a few steps of fine-tuning. These vulnerabilities necessitate new approaches for enabling the safe release of open-weight LLMs. We develop a method, called TAR, for building tamper-resistant safeguards into open-weight LLMs such that adversaries cannot remove the safeguards even after hundreds of steps of fine-tuning. In extensive evaluations and red teaming analyses, we find that our method greatly improves tamper-resistance while preserving benign capabilities. Our results demonstrate that progress on tamper-resistance is possible, opening up a promising new avenue to improve the safety and security of open-weight LLMs.