Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

How to generate {target argument} #1

Open
zggg1p opened this issue Nov 17, 2024 · 1 comment
Open

How to generate {target argument} #1

zggg1p opened this issue Nov 17, 2024 · 1 comment

Comments

@zggg1p
Copy link

zggg1p commented Nov 17, 2024

Excuse me, how is target_argument generated in algorithm 2, and is target_argument replaced with a specific task or is it still represented by a placeholder during the black-box attack?
截屏2024-11-17 13 38 37

@yuanmengqi
Copy link
Collaborator

Thank you for raising these questions!

Regarding the first question: In practical attack scenarios, the attack targets can be flexibly chosen by the attacker. For our experiments, we manually selected some targets different from the original user task target (sourced from the Mind2Web dataset) to showcase the attack's effectiveness. For example, if the user's original task was to buy Apple stocks, we randomly picked other companies, such as Amazon or Microsoft, as the attack targets.

Regarding the second question: Yes, during a black-box attack, the target_argument is replaced with the specific task.

Let us know if you have further questions or need more clarification. We're happy to help!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants