Fast Adversarial Attacks on Language Models In One GPU Minute

Vinu Sankar Sadasivan, Shoumik Saha, Gaurang Sriramanan, Priyatham Kattakinda, Atoosa Chegini, Soheil Feizi
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:42976-42998, 2024.

Abstract

In this paper, we introduce a novel class of fast, beam search-based adversarial attack (BEAST) for Language Models (LMs). BEAST employs interpretable parameters, enabling attackers to balance between attack speed, success rate, and the readability of adversarial prompts. The computational efficiency of BEAST facilitates us to investigate its applications on LMs for jailbreaking, eliciting hallucinations, and privacy attacks. Our gradient-free targeted attack can jailbreak aligned LMs with high attack success rates within one minute. For instance, BEAST can jailbreak Vicuna-7B-v1.5 under one minute with a success rate of 89% when compared to a gradient-based baseline that takes over an hour to achieve 70% success rate using a single Nvidia RTX A6000 48GB GPU. BEAST can also generate adversarial suffixes for successful jailbreaks that can transfer to unseen prompts and unseen models such as GPT-4-Turbo. Additionally, we discover a unique outcome wherein our untargeted attack induces hallucinations in LM chatbots. Through human evaluations, we find that our untargeted attack causes Vicuna-7B-v1.5 to produce $\sim$15% more incorrect outputs when compared to LM outputs in the absence of our attack. We also learn that 22% of the time, BEAST causes Vicuna to generate outputs that are not relevant to the original prompt. Further, we use BEAST to generate adversarial prompts in a few seconds that can boost the performance of existing membership inference attacks for LMs. We believe that our fast attack, BEAST, has the potential to accelerate research in LM security and privacy.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-sadasivan24a, title = {Fast Adversarial Attacks on Language Models In One {GPU} Minute}, author = {Sadasivan, Vinu Sankar and Saha, Shoumik and Sriramanan, Gaurang and Kattakinda, Priyatham and Chegini, Atoosa and Feizi, Soheil}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {42976--42998}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/sadasivan24a/sadasivan24a.pdf}, url = {https://proceedings.mlr.press/v235/sadasivan24a.html}, abstract = {In this paper, we introduce a novel class of fast, beam search-based adversarial attack (BEAST) for Language Models (LMs). BEAST employs interpretable parameters, enabling attackers to balance between attack speed, success rate, and the readability of adversarial prompts. The computational efficiency of BEAST facilitates us to investigate its applications on LMs for jailbreaking, eliciting hallucinations, and privacy attacks. Our gradient-free targeted attack can jailbreak aligned LMs with high attack success rates within one minute. For instance, BEAST can jailbreak Vicuna-7B-v1.5 under one minute with a success rate of 89% when compared to a gradient-based baseline that takes over an hour to achieve 70% success rate using a single Nvidia RTX A6000 48GB GPU. BEAST can also generate adversarial suffixes for successful jailbreaks that can transfer to unseen prompts and unseen models such as GPT-4-Turbo. Additionally, we discover a unique outcome wherein our untargeted attack induces hallucinations in LM chatbots. Through human evaluations, we find that our untargeted attack causes Vicuna-7B-v1.5 to produce $\sim$15% more incorrect outputs when compared to LM outputs in the absence of our attack. We also learn that 22% of the time, BEAST causes Vicuna to generate outputs that are not relevant to the original prompt. Further, we use BEAST to generate adversarial prompts in a few seconds that can boost the performance of existing membership inference attacks for LMs. We believe that our fast attack, BEAST, has the potential to accelerate research in LM security and privacy.} }
Endnote
%0 Conference Paper %T Fast Adversarial Attacks on Language Models In One GPU Minute %A Vinu Sankar Sadasivan %A Shoumik Saha %A Gaurang Sriramanan %A Priyatham Kattakinda %A Atoosa Chegini %A Soheil Feizi %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-sadasivan24a %I PMLR %P 42976--42998 %U https://proceedings.mlr.press/v235/sadasivan24a.html %V 235 %X In this paper, we introduce a novel class of fast, beam search-based adversarial attack (BEAST) for Language Models (LMs). BEAST employs interpretable parameters, enabling attackers to balance between attack speed, success rate, and the readability of adversarial prompts. The computational efficiency of BEAST facilitates us to investigate its applications on LMs for jailbreaking, eliciting hallucinations, and privacy attacks. Our gradient-free targeted attack can jailbreak aligned LMs with high attack success rates within one minute. For instance, BEAST can jailbreak Vicuna-7B-v1.5 under one minute with a success rate of 89% when compared to a gradient-based baseline that takes over an hour to achieve 70% success rate using a single Nvidia RTX A6000 48GB GPU. BEAST can also generate adversarial suffixes for successful jailbreaks that can transfer to unseen prompts and unseen models such as GPT-4-Turbo. Additionally, we discover a unique outcome wherein our untargeted attack induces hallucinations in LM chatbots. Through human evaluations, we find that our untargeted attack causes Vicuna-7B-v1.5 to produce $\sim$15% more incorrect outputs when compared to LM outputs in the absence of our attack. We also learn that 22% of the time, BEAST causes Vicuna to generate outputs that are not relevant to the original prompt. Further, we use BEAST to generate adversarial prompts in a few seconds that can boost the performance of existing membership inference attacks for LMs. We believe that our fast attack, BEAST, has the potential to accelerate research in LM security and privacy.
APA
Sadasivan, V.S., Saha, S., Sriramanan, G., Kattakinda, P., Chegini, A. & Feizi, S.. (2024). Fast Adversarial Attacks on Language Models In One GPU Minute. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:42976-42998 Available from https://proceedings.mlr.press/v235/sadasivan24a.html.

Related Material