arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2412.07097cs.CRcs.AI
收藏

论评估开放权重 LLM 安全防护措施的持久性

On Evaluating the Durability of Safeguards for Open-Weight LLMs

Princeton University(普林斯顿大学);Google(谷歌公司)

Xiangyu Qi, Boyi Wei, Nicholas Carlini, Yangsibo Huang, Tinghao Xie, Luxi He, Matthew Jagielski, Milad Nasr, Prateek Mittal, Peter Henderson

AI总结:

本文通过案例研究指出,开放权重 LLM 的持久安全防护在评估上存在易夸大防护效果的陷阱,并主张以严格界定的威胁模型审慎评估其局限。

摘要:

从模型开发者到政策制定者,各利益相关方都力求将大语言模型(LLM)的双重用途风险降至最低。实现这一目标面临的一个公开挑战是:即使模型可通过微调进行定制,或者模型权重完全开放,技术防护措施是否仍能阻止 LLM 被滥用。作为回应,近期多项研究提出了为开放权重 LLM 构建持久安全防护的方法,使其能够抵御通过微调对模型权重实施的对抗性修改。这有望在攻击者可直接微调模型权重的强威胁模型下,提高攻击者的成本。然而,本文主张应更谨慎地刻画这些方法的局限。通过若干案例研究,我们证明,即使只是评估这些防御也极其困难,而且很容易误导受众,使其认为安全防护比实际情况更持久。我们从所识别的评估陷阱中吸取教训,并建议未来研究谨慎地将主张限定在更受约束、定义更明确且经过严格检验的威胁模型内,从而为利益相关方提供更有用、更坦诚的评估。

英文摘要:

Stakeholders -- from model developers to policymakers -- seek to minimize the dual-use risks of large language models (LLMs). An open challenge to this goal is whether technical safeguards can impede the misuse of LLMs, even when models are customizable via fine-tuning or when model weights are fully open. In response, several recent studies have proposed methods to produce durable LLM safeguards for open-weight LLMs that can withstand adversarial modifications of the model's weights via fine-tuning. This holds the promise of raising adversaries' costs even under strong threat models where adversaries can directly fine-tune model weights. However, in this paper, we urge for more careful characterization of the limits of these approaches. Through several case studies, we demonstrate that even evaluating these defenses is exceedingly difficult and can easily mislead audiences into thinking that safeguards are more durable than they really are. We draw lessons from the evaluation pitfalls that we identify and suggest future research carefully cabin claims to more constrained, well-defined, and rigorously examined threat models, which can provide more useful and candid assessments to stakeholders.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑