Repository navigation
Grill #1484: process Eval suite decisions, ADRs 0027–0028 and research - #1890
Merged
Merged
Conversation
Owner
Author
|
Execution evidence: |
Merged
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Documentation only: the #1484 grill, the decisions it reached, and the research behind them. Specs #1484 and #1872 cite these files.
GLOSSARY.md # Arm names a Process; Trial may be many sessions; # Bar fixes who grades; new term: Grader panel docs/adr/ +├── 0027-a-trial-carries-out-a-process-but-is-not-a-command.md +└── 0028-who-decides-a-verdict-is-fixed-in-the-bar.md docs/grill-with-docs/done/ +└── 1484-implement-suite.md # 10 rounds, Q1.1–Q10.2 docs/research/ +├── eval-process-arms-and-human-review.md # Langfuse et al., external runners, blind review +└── eval-grader-panel.md # same-family Graders, panel size, Kimi/DeepSeek HarnessesWhat the grill decided, and where it was published:
Evidence
Execution evidence
No code changes.
scripts/affected-tests origin/mainselects the wholecargo testsuite (the glossary changed). thirdshift is Linux-only, so it runs in the Lima VM. The result is in a comment below, and CI runs it too.Merge Danger
Door: two-way
Revert the commit. The Specs would then cite missing files.
Blast Radius: documentation
The glossary wording for Arm, Trial and Bar changes for every agent that reads it. The new wording describes behaviour #1484 and #1872 will build, so today's code doesn't yet match the Arm's Process clause, a multi-session Trial or the Bar's grading settings.