AI foundry - New built-in role to peform "human-in-the-loop" review of evals outcome so that security evals outcome/data is protected using least privilegeby

**Is your feature request related to a problem? Please describe.**
Currently anyone having access to AI Foundry priject can view security evals outcome. howver, these might be sensitive so as to restrict only few people in an org to view and analyse it. There is "BLUR" button but looks like anyone with access to AI Foundry project can toggle it to see the evals outcome.

This is an issue because if evals are run for prompt injection attacks or code vulnerabilties then anyone who is having access to AI Foundry project can view the "bad prompts" and might use it to harm organisation (insider threat).

Ideally, access to evals results has to be controlled via separate role.

**Describe the solution you'd like**
New built in role to allow view of evals results

**Describe alternatives you've considered**
No alternative currently

**Additional context**