-
-
Notifications
You must be signed in to change notification settings - Fork 378
Improve reporting on robustness/ethical issues [GSK-1765] #1411
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
Conversation
mattbit
commented
Sep 20, 2023
- Remove percentage and put x our of N samples changed because we don't know how many samples contain countries
- Add the probabilities next to the prediction
GSK-1765 Enhance `BaseTextPerturbationDetector` outputs
|
| "feature": feature, | ||
| "domain": f"Feature `{feature}`", | ||
| "deviation": f"{round(fail_rate * 100, 2)}% of tested samples changed prediction after perturbation", | ||
| "deviation": f"{failed_size} out of {slice_size} tested samples ({round(fail_rate * 100, 2)}%) changed prediction after perturbation", |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we should unify the convention here, I used {failed_size}/{slice_size} in the prompt-injection.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Let’s keep what is the most clear in your opinion
|
@rabah-khalek what should we do with this? It’s been a while |
|
sorry, it's good to go once the conflicts are resolved. |
Co-authored-by: Rabah Abdul Khalek <[email protected]>
|
Taking care of this now |
|
Kudos, SonarCloud Quality Gate passed! |







