Meta-eval: (6/7) Measure held-out Elo error against annotation budget - #106
Open
ErlisLushtaku wants to merge 3 commits into
Open
Meta-eval: (6/7) Measure held-out Elo error against annotation budget#106ErlisLushtaku wants to merge 3 commits into
ErlisLushtaku wants to merge 3 commits into
Conversation
Hybrid Bradley-Terry (other models' human battles plus K LLM labels for the held-out model) is compared to full-human BT at each --meta_eval.elo_gap_battles budget. Ties are dropped after sampling so the budget stays the number of purchased annotations. Ranking and Elo-gap fits still use one forward-order row per battle.
This was referenced Aug 15, 2026
# Conflicts: # README.md
ErlisLushtaku
changed the title
ErlisLushtaku
marked this pull request as ready for review
August 17, 2026 09:21
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
Note
Packaged meta-eval split from closed PR #76. Meta-eval does not use the unified
do_inferencecache, and the cache stack (#94-#100) does not wire it up either: caching will be added to meta-eval once those PRs are merged.Held-out Elo MAE compares a hybrid Bradley-Terry fit (other models' human battles plus K LLM labels for the held-out model) to full-human BT at each
--meta_eval.elo_gap_battlesbudget. Ties are dropped after sampling so the budget stays the number of purchased annotations.This is stacked on #105.