Repository navigation

Fix Tool Call Success scoring, binary_aggregate and Decision input order, and name Decision failure reasons - #49804

Merged

bonanza89 merged 19 commits into

Azure:mainfrom

bonanza89:users/aqardahji/token-reuse-and-tool-call-success

Nov 4, 2026

Merged

Fix Tool Call Success scoring, binary_aggregate and Decision input order, and name Decision failure reasons#49804
bonanza89 merged 19 commits into
Azure:mainfrom
bonanza89:users/aqardahji/token-reuse-and-tool-call-success

Commits on Nov 3, 2026
- Reuse credential tokens across model requests and fix Tool Call Success scoring
  Ahmad Qardahji
  committed
  4c96b29
- Merge branch 'main' into users/aqardahji/token-reuse-and-tool-call-success
  bonanza89
  authored
  dc10282
- Share the chat token scope with Decision requests and drop the redundant retry list
  Ahmad Qardahji
  committed
  5c068e8
- Name the reason a failed Decision request was refused, and share only authentication failures with waiting requests
  Ahmad Qardahji
  committed
  3881048
- Keep Decision on the Cognitive Services scope and keep token failures within each kind of request
  Ahmad Qardahji
  committed
  8ac9c3a
- Format the Decision error-kind patterns and tests with black
  Ahmad Qardahji
  committed
  84df6b0
- Format the Decision batch persistence test with black
  Ahmad Qardahji
  committed
  44097f4
- Ask credentials that also implement the synchronous context manager protocol on a worker thread
  Ahmad Qardahji
  committed
  d960c79
- Move token reuse to its own pull request
  Ahmad Qardahji
  committed
  4d7a193
- Merge branch 'main' into users/aqardahji/token-reuse-and-tool-call-success
  bonanza89
  authored
  773118a
- Merge branch 'main' into users/aqardahji/token-reuse-and-tool-call-success
  bonanza89
  authored
  aa69a25
- Treat empty tool call fields as absent in Tool Call Success
  Ahmad Qardahji
  committed
  9e48c79
- Count only passed and failed rows in binary_aggregate
  Ahmad Qardahji
  committed
  4fe6316
Commits on Nov 4, 2026
- Keep evaluator inputs in their declared order
  Ahmad Qardahji
  committed
  8b8fa38
- Keep the last result column deciding binary_aggregate
  Ahmad Qardahji
  committed
  abc053b
- Run the hash seed test's processes from an empty directory
  Ahmad Qardahji
  committed
  70135d2
- Merge the main updates on the PR branch
  Ahmad Qardahji
  committed
  d100061
- Merge branch 'main' into users/aqardahji/token-reuse-and-tool-call-success
  bonanza89
  authored
  b567723
- Leave unsupported role and content type values to the judge
  Ahmad Qardahji
  committed
  0028aa3

Commits

Commits on Nov 3, 2026

Commits on Nov 4, 2026