docs: Add GEPA prompt optimization and human agreement specs#95
Open
vivian-xie-db wants to merge 2 commits into
Open
docs: Add GEPA prompt optimization and human agreement specs#95vivian-xie-db wants to merge 2 commits into
vivian-xie-db wants to merge 2 commits into
Conversation
Add declarative specifications for two key features: - PROMPT_OPTIMIZATION_SPEC.md: GEPA optimizer pipeline, training data, score normalization, predict_fn, config persistence, auto-reconnect - HUMAN_AGREEMENT_SPEC.md: GDPVal A^HH human-to-human agreement metric, rating normalization, pairwise agreement %, IRR integration Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
|
IRR metrics are currently crammed into the Judge evaluation spec where they probably shouldn't be. We should move the IRR section out and merge it into the human agreement spec you've added here |
FMurray
reviewed
Feb 13, 2026
| |--------|-----------------|-----------------| | ||
| | **GDPVal A^HH** (this spec) | Human vs Human agreement | IRR Results page | | ||
| | Pairwise Agreement % | Human vs Human agreement (percentage) | IRR Results page | | ||
| | Cohen's Kappa | Judge vs Human agreement | Judge Tuning page | |
There was a problem hiding this comment.
GDPval uses A^HA as well? Why Cohen's Kappa still?
Collaborator
Author
There was a problem hiding this comment.
Cohen's Kappa is in judge tuning page. There is a small section above evaluation results where it is showing the Cohen's Kappa score, which has already been there since version 1.0
…el in judge evaluation spec Remove the generic IRR section (Krippendorff's Alpha, Cohen's Kappa for rater pairs) from JUDGE_EVALUATION_SPEC and replace with a detailed Cohen's Kappa Metrics Panel spec covering the judge-vs-human agreement metrics displayed after evaluation on the Judge Tuning page. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
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
2 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.
Summary
optimize_promptsAPI, training data from annotated traces, score normalization,predict_fnbehavior, custom endpoint support, config persistence, auto-reconnect, and score improvement display.E[1 - |H_1 - H_2|], rating normalization (Likert/binary), pairwise agreement %, per-metric breakdowns, and IRR integration.Test plan
🤖 Generated with Claude Code