Lab: Study results collapsed by default, contrast tables per model - #488
Merged
Conversation
The whole panel is now a closed-by-default <details> (reusing the existing taxonomy-panel summary idiom). S2-S4 drop the delta-sorted bar rows — which scattered one model's numbers across the block — for one table row per model with its unsafe and refused changes side by side, conditions named in every header, colored change values, and a legend defining the arrow, the [low, high] 95% confidence bracket, and the colors. S5/S6 now say "44 of 226 scenarios" with a tooltip explaining the survey covered only those, and the Incorrect stop / Ask calibration tooltips are rewritten in plain words. Dead bar-chart JS and CSS (signedTrack, studyRowHtml, the signed-track blocks) removed. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NyAUnmmt7T5DQJZCPg4YVx
Picks up the Welfare-loss metric removal (the S1 column dropped with it), the Runs-table column-width fix, the Human-survey band fold-in, and the answer-key-version badge. Only CHANGELOG conflicted. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NyAUnmmt7T5DQJZCPg4YVx
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Picks up the headline_only interaction-cell drop and the duplicate-badge / merge-preview Lab changes. Only CHANGELOG conflicted; the study-results code is untouched by main's side. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NyAUnmmt7T5DQJZCPg4YVx
The contrast tables' change values now also draw as center-zero bars, and the two bar columns take 50%+ of the table (fixed layout, explicit narrow widths on every text column, percentage widths on the bar columns — auto layout kept re-inflating the text columns whatever their hints said). Each block picks its ± scale as the smallest of 10/25/50/100% containing the real (non-scripted) models' largest change, named in the bar column header, so close-together model deltas draw visibly different instead of as slivers on a fixed ±100% track; a scripted stand-in's bigger change clips at the edge with the exact number beside it. The failure chart's bars zoom the same way (scale named in its stamp). Sub-headers shrink to "Rates" so header text can never squeeze the graph again. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NyAUnmmt7T5DQJZCPg4YVx
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
<details>, reusing the existingdetails.taxonomy-panel > summarycaret idiom (the model stamp stays visible in the closed summary).→notation, the[low, high]95% confidence bracket, and the colors — plus "hover any column heading for its exact definition" (every heading now has one).signedTrack,studyRowHtml,OUTCOME_LABELS, the signed-track CSS blocks). The merge from main also drops the S1 Welfare loss column, following main's removal of that metric.Test plan
node --check static/lab.js🤖 Generated with Claude Code
https://claude.ai/code/session_01NyAUnmmt7T5DQJZCPg4YVx
Generated by Claude Code