-
Notifications
You must be signed in to change notification settings - Fork 4
Open
Description
I am evaluating my own agent on MLA-Trust. I don't know how to interpret the result.xlsx produced by running bash scripts/web/eval.sh.
Particularly, I want to know how to compute RtE and ASR from the result.xlsx.
RtE = # refuse_answer / # total_number
ASR = # is_success_attack / # total_number
Is it correct?
Thanks for your time!
Metadata
Metadata
Assignees
Labels
No labels