-
Notifications
You must be signed in to change notification settings - Fork 450
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
llamagaurd3 identifies code as 'violent crime' #63
Comments
Hi @visagansanthanam-unisys can you give us other examples of this? No, this is not expected behavior |
@EricMichaelSmith here are some more examples |
Hi @visagansanthanam-unisys the two models (Llama Guard 1B and 8B) are different in terms of training data and underlying base models. Specifically, the 1B model was not trained on the coding interpreter category, which can lead to limitations for code input. |
I am trying to have llamagaurd3 for a use case and I see that the model identifies any source code as unsafe violent crime. is this is a expected behavior
The text was updated successfully, but these errors were encountered: