Add configurable adversarial benchmark pipeline - #2436
Open
hannahwestra25 wants to merge 13 commits into
Open
Conversation
hannahwestra25
marked this pull request as draft
August 19, 2026 22:00
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 5b5e066d-8e77-4ba9-8fc8-56f4745ef365
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 5b5e066d-8e77-4ba9-8fc8-56f4745ef365
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 5b5e066d-8e77-4ba9-8fc8-56f4745ef365
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 5b5e066d-8e77-4ba9-8fc8-56f4745ef365
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 5b5e066d-8e77-4ba9-8fc8-56f4745ef365
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 5b5e066d-8e77-4ba9-8fc8-56f4745ef365
hannahwestra25
force-pushed
the
hannahwestra25-adversarial-benchmark-system-prompt
branch
from
August 20, 2026 16:34
b2cbedd to
7f7b98a
Compare
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 5b5e066d-8e77-4ba9-8fc8-56f4745ef365
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 5b5e066d-8e77-4ba9-8fc8-56f4745ef365
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 5b5e066d-8e77-4ba9-8fc8-56f4745ef365
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 5b5e066d-8e77-4ba9-8fc8-56f4745ef365
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 5b5e066d-8e77-4ba9-8fc8-56f4745ef365
hannahwestra25
changed the title
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 5b5e066d-8e77-4ba9-8fc8-56f4745ef365
hannahwestra25
marked this pull request as ready for review
August 20, 2026 20:26
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
1 participant
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.
Summary
Adds an ADO script to run the adversarial benchmark scenario so we can automatically run the scenario as new models are released. Keeps targets, dataset, sample size, concurrency, retries, and log level configurable.
The pipeline runs exactly the space-separated technique names supplied through one
techniquesparameter; there are no candidate/default/union presets.Right now it uses a deterministic 14-objective HarmBench subset and reuses the existing integration-test service connection and Key Vault environment. It retains phase and exit diagnostics, logs, SQLite state, partial or completed result views, and retry-deduplicated per-technique metrics as pipeline artifacts.
Notes