{"ok":true,"scope":"public","version":"agentbench-v1","atlasLabel":"atlas-agent-pipeline","baselineLabel":"frontier-llm-agent (synthetic baseline)","n":12,"chainTip":"4bde314488356334ca0363b7de8a3d7213c3adeb26083a997b218000b3069b8a","genesisHash":"","kindBuckets":[{"n":4,"atlas":{"successRate":1,"avgScore":1,"avgFaithfulness":1},"baseline":{"successRate":0,"avgScore":0.584,"avgFaithfulness":0.584},"delta":{"successRate":1,"avgScore":0.416,"avgFaithfulness":0.416,"atlasWins":4,"baselineWins":0,"ties":0,"winRate":1},"kind":"build-a-model"},{"n":4,"atlas":{"successRate":1,"avgScore":1,"avgFaithfulness":1},"baseline":{"successRate":0,"avgScore":0.729,"avgFaithfulness":0.729},"delta":{"successRate":1,"avgScore":0.271,"avgFaithfulness":0.271,"atlasWins":4,"baselineWins":0,"ties":0,"winRate":1},"kind":"multi-hop"},{"n":4,"atlas":{"successRate":1,"avgScore":1,"avgFaithfulness":1,"catchRate":1},"baseline":{"successRate":0.25,"avgScore":0.25,"avgFaithfulness":0.25,"catchRate":0.25},"delta":{"successRate":0.75,"avgScore":0.75,"avgFaithfulness":0.75,"atlasWins":3,"baselineWins":0,"ties":1,"winRate":0.75,"catchRate":0.75},"kind":"self-falsification"}],"regimeBuckets":[{"n":2,"atlas":{"successRate":1,"avgScore":1,"avgFaithfulness":1,"catchRate":1},"baseline":{"successRate":0,"avgScore":0.334,"avgFaithfulness":0.334,"catchRate":0},"delta":{"successRate":1,"avgScore":0.666,"avgFaithfulness":0.666,"atlasWins":2,"baselineWins":0,"ties":0,"winRate":1,"catchRate":1},"regime":"CREDIT"},{"n":2,"atlas":{"successRate":1,"avgScore":1,"avgFaithfulness":1,"catchRate":1},"baseline":{"successRate":0,"avgScore":0.375,"avgFaithfulness":0.375,"catchRate":0},"delta":{"successRate":1,"avgScore":0.625,"avgFaithfulness":0.625,"atlasWins":2,"baselineWins":0,"ties":0,"winRate":1,"catchRate":1},"regime":"MACRO"},{"n":2,"atlas":{"successRate":1,"avgScore":1,"avgFaithfulness":1,"catchRate":1},"baseline":{"successRate":0,"avgScore":0.375,"avgFaithfulness":0.375,"catchRate":0},"delta":{"successRate":1,"avgScore":0.625,"avgFaithfulness":0.625,"atlasWins":2,"baselineWins":0,"ties":0,"winRate":1,"catchRate":1},"regime":"RISK"},{"n":1,"atlas":{"successRate":1,"avgScore":1,"avgFaithfulness":1},"baseline":{"successRate":0,"avgScore":0.667,"avgFaithfulness":0.667},"delta":{"successRate":1,"avgScore":0.333,"avgFaithfulness":0.333,"atlasWins":1,"baselineWins":0,"ties":0,"winRate":1},"regime":"VOL"},{"n":1,"atlas":{"successRate":1,"avgScore":1,"avgFaithfulness":1},"baseline":{"successRate":0,"avgScore":0.5,"avgFaithfulness":0.5},"delta":{"successRate":1,"avgScore":0.5,"avgFaithfulness":0.5,"atlasWins":1,"baselineWins":0,"ties":0,"winRate":1},"regime":"WEALTH"},{"n":4,"atlas":{"successRate":1,"avgScore":1,"avgFaithfulness":1,"catchRate":1},"baseline":{"successRate":0.25,"avgScore":0.729,"avgFaithfulness":0.729,"catchRate":1},"delta":{"successRate":0.75,"avgScore":0.271,"avgFaithfulness":0.271,"atlasWins":3,"baselineWins":0,"ties":1,"winRate":0.75,"catchRate":0},"regime":"EQUITY"}],"overall":{"n":12,"atlas":{"successRate":1,"avgScore":1,"avgFaithfulness":1,"catchRate":1},"baseline":{"successRate":0.083,"avgScore":0.521,"avgFaithfulness":0.521,"catchRate":0.25},"delta":{"successRate":0.917,"avgScore":0.479,"avgFaithfulness":0.479,"atlasWins":11,"baselineWins":0,"ties":1,"winRate":0.917,"catchRate":0.75}},"createdAt":null,"computed":true,"synthetic":true,"note":"Atlas agent pipeline (orchestration + reconciliation + self-falsification) vs a synthetic frontier-agent baseline on the SAME long-horizon agent tasks: build-a-reconciled-model, multi-hop research, and self-falsification catch-rate. Baseline is a deterministic synthetic stand-in (no live frontier key) — every cell is synthetic and the whole leaderboard is sha256-chained + independently re-derivable at /verify."}