Test Sets / Match Runs

Browse every persisted match run for a team — human-driven Lab searches and agent-generated test sets alike — re-run one pinned to a chosen algorithm version, and compare two runs side by side to judge which version performed better.