Open
Conversation
rlundeen2
reviewed
Apr 15, 2026
rlundeen2
reviewed
Apr 15, 2026
jbolor21
reviewed
Apr 16, 2026
jbolor21
reviewed
Apr 16, 2026
rlundeen2
reviewed
Apr 16, 2026
| ## Viewing Scoring Metrics | ||
|
|
||
| There are a few different ways to view metrics for specific scoring configurations. | ||
|
|
Contributor
There was a problem hiding this comment.
Can you link to the docs here?
rlundeen2
reviewed
Apr 16, 2026
| metrics = await my_scorer.evaluate_async(num_scorer_trials=3) | ||
| ``` | ||
|
|
||
| The framework checks the JSONL registry for an existing entry matching the scorer's evaluation hash before running. It only re-runs the evaluation if no entry exists, the dataset version changed, the harm definition version changed, or the requested number of trials exceeds what's stored. You can skip this registry check entirely with `update_registry_behavior=RegistryUpdateBehavior.NEVER_UPDATE` if you're experimenting and don't want to write to the registry. (This should most often be the case since metrics saved to the registry are managed directly by Microsoft's AI Red Team. Please don't hesitate to reach out however if you'd like to add metrics for new scoring configurations to our registries.) The below shows a simple example of running an evaluation: |
Contributor
There was a problem hiding this comment.
I would take this further because I think this could be a hook for people who have never used pyrit.
Do you want to see how accurate your judge is compared to pyrit or anything else? We collected human responses and have a framework you can use, just adapt your judge or evaluation into a pyrit scorer and run it.
rlundeen2
approved these changes
Apr 16, 2026
Contributor
rlundeen2
left a comment
rlundeen2
left a comment
There was a problem hiding this comment.
I'd wait for one more approval here, but looks good! I have two notes :)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
This PR adds a blog documenting our scorer evaluation background, story, and process!
Tests and Documentation
N/A