Results autoresearch
Per-iteration ledger from results.tsv (12 columns, RESULTS_TSV_HEADER in autoresearch/train.py:1284). Each row pairs with one results.jsonl line for the full per-dim breakdown. Fitness trend rendered inline as a Unicode block sparkline.
Trend · fitness sparkline
fitness — no iterations recorded
Results · 0 rows
| session_id | gen_tag | commit | fitness | critical_min | critical_mean | auxiliary_mean | stability_score | info_mean | dim_count_engaged | verdict | description |
|---|---|---|---|---|---|---|---|---|---|---|---|
| No iterations recorded yet. | |||||||||||
Source: autoresearch/state/results.tsv + autoresearch/state/results.jsonl (12-col header per RESULTS_TSV_HEADER, train.py:1284).
Published by .github/workflows/pages.yml on every main push.
Repo: github.com/mangowhoiscloud/geode
Harness chip legend: PAYGAPI key billing · Claude CodeMax OAuth · CodexChatGPT OAuth · GEODEself-target wrapper.
Rendered against GEODE v0.99.92 · DESIGN.md schema 1 · built 2026-05-30 09:11.
Dim subset: 22 (geode_5axes). Pipeline phases: 8. Baseline schema: v2 (PR-2).