Process dump — numbered artifacts, one-line captions. Not a marketing page.
Engine: kvnloo/evolution-lab ·
Claims: frontier-kb ·
Runtime: openjev
Numpy mushroom-body student on typed Hermes recovery. Frozen PN→KC, plastic KC→MBON only. Not FlyGym, not OpenEvolve, not a Qwen swap. SearchDriver ≠ frozen judge.
| Family | Confirm | Val | Params |
|---|---|---|---|
| teacher / rule | 1.00 | 1.00 | 0 |
| direct_input | 0.79 | 0.63 | 75 |
| mlp | 1.00 | 1.00 | 4032 |
| gru | 0.92 | 0.90 | 3000 |
| fixed_reservoir | 1.00 | 0.96 | 3264 |
| local_plasticity | 1.00 | 0.98 | 640 |
Kill criterion did not fire. MaleCNS SGD still refused. Encoder is flatten+last+maxpool Hermes history, not the compound eye.
lab / MAP-Elites — discover families on the control table.
Not the production kernel tuner.autoresearch — 1-knob champion chase; frozen bench.py / select.py.
Hard gates: closed-loop ≥ 1.0, confirm/val ≥ 0.95, ood ≥ 0.85, violations = 0.cycle — ABAB A-timescale: 1-knob session until plateau, then sweep. World in league/abab.json.
B is still frozen-judge autoresearch/sweep. Daemon opt-in (--forever).sweep — 75% of logical CPUs, OMP_NUM_THREADS=1.
Contended latency is not comparable to serial league. Serial rebench before keep.jev-distill — fly pre-predicts TypeSafe Jev skill routing from OMP history × Jev logs.
Distillation, not RLHF. Jev already emits typed choices.| Start | League | |
|---|---|---|
| id | seed=1 / 640 w | fly-4080a10a |
| knobs | hidden=128, dagger=0 | hidden=96, dagger=3 |
| latency_score | 12.20 → 6.32 | 5.923 |
| closed-loop | 1.0 | 1.0 |
| advise p99 | — | 0.13 ms vs Jev ~272 ms |
OMP bundle: data/p0/recovery_student.npz. Cycle stopped on operator request after 30 sessions / 23 idle.
.hermes/abab/fly-cycle-meta.json (local).72 OMP history.db prompts time-joined to Jev route logs (median 0.88 s). 7-way confirm: fly = majority = 0.33. Ridge 0.17. Fly 672 w, 0.028 ms. Cannot beat Jev without an outcome verifier.
k_winners {5,8,12} failed the closed-loop gate. k_winners=10 (the 10% heuristic) holds. Best contended score 6.31 still worse than serial 5.92. Control rebenched at 9.24 under load — do not keep from contended benches.
Serial rebench of top rows: none beat 5.923. Champion stands.
python -m evolution_lab lab --evolve-generations 2 --dagger-rounds 2
python -m evolution_lab autoresearch --tag YYYYMMDD --max-experiments 20
python -m evolution_lab cycle --forever --skip-unit-tests
python -m evolution_lab sweep --workers 15
python -m evolution_lab jev-distill
python -m evolution_lab recovery '{"action":"status"}'
Maintainer roadmap stays ROADMAP.md. Workers do not rewrite it.