WIP: (1/7) Declare meta-evaluation as a packaged task - #101
Draft
ErlisLushtaku wants to merge 1 commit into
Draft
Conversation
Meta-evaluation scores a judge against human arena votes. Registering it as a task definition rather than a separate entrypoint means it inherits task discovery, pinned dataset revisions, and language variants for free. The protocol declares no generation and no baseline, since both completions already exist in the arena, and no scoring adapter yet; the agreement and ranking metrics arrive in later PRs of this stack. Sampling keeps the most-battled models and draws battles per model, which is what makes the annotations usable for a ranking fit later on. Language selection moves out of the ELO runner into a shared helper, and the arena turn extractor becomes public under a name that reflects that it reads assistant turns too.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
WIP: packaged meta-eval split from closed PR #76. Caching is not in this stack; it stays with the unified
do_inferencecache PRs.Adds
meta-eval-lmarena-140kandmeta-eval-compariaas packaged YAML tasks on the same arena battle adapter as ELO, plus deterministic top-model / per-model sampling.There is no generation and no
model.name. Completions already exist in the arena.This is stacked on #93.
Stack: