Experiment Lab

Automated A/B testing, quality tracking, and config optimization

/100
System Quality Score (7d)
Loading...

Tier 1 — Atomic

loading...

Tier 2 — Composite

loading...

Tier 3 — Evolution

loading...
Active Experiments
No active experiments
Recent Results
ExperimentAgentTierVerdictDeltaCompleted
Loading...
Experiment Presets
Leaderboard
No completed experiments yet

Architecture matrix (α–θ): pipeline personalities and config overrides

Loading architectures…

MAP-Elites archive for the selected agent (use Overview filters)

Grid coverage
Occupied / total
Best score
Worst score
Mean score
Click “Suggest next experiment” to load a recommendation.
Best configs
Loading…

Quality stagnation, trends, and breakout suggestions (T5 composite)

Quality timeline (30d daily mean)

Portable wins and cross-agent generalization

Loading portable wins…