Skip to content

feat(plugin): add trial output-consistency metric via embedding similarity#375

Merged
christso merged 1 commit into
mainfrom
feat/368-trial-consistency-metric
Feb 25, 2026
Merged

feat(plugin): add trial output-consistency metric via embedding similarity#375
christso merged 1 commit into
mainfrom
feat/368-trial-consistency-metric

Conversation

@christso
Copy link
Copy Markdown
Collaborator

Closes #368

Part of #371 roadmap.

Changes

  • Added examples/features/trial-output-consistency/ with judge, eval, and docs
  • Embedding-based pairwise cosine similarity for trial consistency
  • Handles edge cases (0 trials, 1 trial, 2+ trials)
  • Exposes consistency as a named metric in eval workflows

…arity

Closes #368

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
@christso christso merged commit 2a05b00 into main Feb 25, 2026
1 check was pending
@christso christso deleted the feat/368-trial-consistency-metric branch February 25, 2026 12:44
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

feat(plugin): trial output-consistency metric via embedding similarity

1 participant