Request: AppBench evaluation licence, official harness, and grading route
Hello AfterQuery team,
I am preparing a reproducible academic evaluation using the public
AfterQuery/App-Bench dataset pinned at revision
de80d5bcd404adee5307311571e512b5c37e6112. The dataset currently exposes
only AppBench vExternal.csv: its dataset card is null and the revision has no
licence, terms, or notice file.
Could you point me to the official channel for the following, or provide them
here if public?
- Written permission or a licence for evaluation use and publication of
aggregate results.
- The official deployment/reset runner or a provider-hosted evaluation path,
including artifact and side-effect verification.
- The required two-grader protocol, re-adjudication rules, and whether a
provider-run grader service is available (including terms and price).
- The authoritative held-out/contamination policy for the six public tasks.
I will not substitute a local protocol, appoint graders, or report a benchmark
score until these requirements have immutable evidence. I am not requesting
credentials or initiating a paid evaluation through this issue.
The same request is open in the dataset discussion:
https://huggingface.co/datasets/AfterQuery/App-Bench/discussions/2
Thank you.
Request: AppBench evaluation licence, official harness, and grading route
Hello AfterQuery team,
I am preparing a reproducible academic evaluation using the public
AfterQuery/App-Benchdataset pinned at revisionde80d5bcd404adee5307311571e512b5c37e6112. The dataset currently exposesonly
AppBench vExternal.csv: its dataset card is null and the revision has nolicence, terms, or notice file.
Could you point me to the official channel for the following, or provide them
here if public?
aggregate results.
including artifact and side-effect verification.
provider-run grader service is available (including terms and price).
I will not substitute a local protocol, appoint graders, or report a benchmark
score until these requirements have immutable evidence. I am not requesting
credentials or initiating a paid evaluation through this issue.
The same request is open in the dataset discussion:
https://huggingface.co/datasets/AfterQuery/App-Bench/discussions/2
Thank you.