feat: autoresearch harness for SNAG optimization#26
Merged
seanfromthepast merged 1 commit intomainfrom Mar 16, 2026
Merged
Conversation
Karpathy-style autoresearch loop that mutates Hydra config parameters across 27 dimensions (6 mechanism clusters), runs simulations, extracts quality metrics (Causal Resolution, coherence, plausibility), and identifies Pareto-optimal configs via quality vs cost tradeoffs. Supports --dry-run mode with deterministic synthetic metrics for testing the mutation/selection loop without API calls.
seanfromthepast
approved these changes
Mar 16, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Adds
autoresearch/directory with Karpathy-style optimization loop for Pro simulation parameters.pro_autoresearch.py— main loop: mutate config → run → evaluate → keep/discardconfig_space.py— 27 mutable dimensions across 6 mechanism clustersmetrics.py— Causal Resolution metric, dry-run synthetic metricspareto.py— Pareto frontier analysisPurely additive — no existing files modified. Safe to merge.
Merge intent
This is the base branch. Once merged, the 8 cluster branches below can PR their result files in:
autoresearch/pro/fidelity→ results + issue Autoresearch Pro-1: Fidelity optimization findings (M1/M2/M5/M6) #18autoresearch/pro/temporal→ results + issue Autoresearch Pro-2: Temporal mode optimization findings (M17) #21autoresearch/pro/knowledge→ results + issue Autoresearch Pro-3: Knowledge provenance findings (M3/M4/M19) #22autoresearch/pro/entities→ results + issue Autoresearch Pro-4: Entity simulation findings (M9-M16) #19autoresearch/pro/models→ results + issue Autoresearch Pro-5: Model routing optimization findings (M18) #20autoresearch/pro/dialog→ results + issue Autoresearch Pro-6: Dialog quality findings (M10/M11) #23autoresearch/pro/generalize→ results + issue Autoresearch Pro-7: Cross-template generalization findings #25autoresearch/pro/tdf-training→ results + issue Autoresearch Pro-8: TDF + training data quality findings #24DO NOT MERGE:
feat/pro/pytorch-backend— modifiestensors.py, needs venv testing first.Test plan
python3 -m autoresearch.pro_autoresearch --dry-run --iterations 10completes <1s