An Adversarial Perspective on Machine Unlearning for AI Safety

September 26, 2024 ยท Declared Dead ยท ๐Ÿ› Trans. Mach. Learn. Res.

๐Ÿ‘ป CAUSE OF DEATH: Ghosted
No code link whatsoever

"No code URL or promise found in abstract"

Evidence collected by the PWNC Scanner

Authors Jakub ลucki, Boyi Wei, Yangsibo Huang, Peter Henderson, Florian Tramรจr, Javier Rando arXiv ID 2409.18025 Category cs.LG: Machine Learning Cross-listed cs.AI, cs.CL, cs.CR Citations 90 Venue Trans. Mach. Learn. Res. Last Checked 4 months ago
Abstract
Large language models are finetuned to refuse questions about hazardous knowledge, but these protections can often be bypassed. Unlearning methods aim at completely removing hazardous capabilities from models and make them inaccessible to adversaries. This work challenges the fundamental differences between unlearning and traditional safety post-training from an adversarial perspective. We demonstrate that existing jailbreak methods, previously reported as ineffective against unlearning, can be successful when applied carefully. Furthermore, we develop a variety of adaptive methods that recover most supposedly unlearned capabilities. For instance, we show that finetuning on 10 unrelated examples or removing specific directions in the activation space can recover most hazardous capabilities for models edited with RMU, a state-of-the-art unlearning method. Our findings challenge the robustness of current unlearning approaches and question their advantages over safety training.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

๐Ÿ“œ Similar Papers

In the same crypt โ€” Machine Learning

Died the same way โ€” ๐Ÿ‘ป Ghosted