Proposal: ensemble + judge-refinement loop for early pipeline tasks#395
Merged
neoneye merged 1 commit intoPlanExeOrg:mainfrom Mar 28, 2026
Merged
Conversation
…tasks Three-stage design per Simon's direction: 1. N non-reasoning candidates run in parallel 2. Reasoning model judges all N, scores + gap hints (short output) 3. Conditional retry with gap hints if below threshold Applies highest-leverage to early tasks (PremiseAttack, RedlineGate, ProjectPlan) where quality propagates downstream. Extends existing LLMExecutor.max_validation_retries pattern to quality-based retries. Backward-compatible with local/Ollama setups.
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
3 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Three-stage ensemble + judge-refinement loop for early pipeline tasks.
Targets PremiseAttackTask and RedlineGateTask first — where quality propagates most downstream. Extends existing LLMExecutor.max_validation_retries pattern to quality-based retries. Backward-compatible with local/Ollama (sequential, no judge step if no judge model configured).
See docs/proposals/ensemble-judge-refinement.md for full design + open questions.
Follows PR #393 (parallel racing proposal).