Maple Brain Lab
A transparent status page for Hermon training, evaluation, and deployment. This page shows the current model lanes, public eval status, latest trained adapters, and the next improvement targets without publishing private machine details.
Average public contract score: 68
Training Policy
Maple Brain Lab is organized around eval-directed improvement. A domain is not trained simply because the loop can run. It is trained because a public or private probe identified missing contract fields, weak refusal quality, poor domain language, or a deployment gap.
Model Lanes
Hermon is the model family. MapleAI is the runtime and OS layer that authorizes actions, records receipts, and controls deployment. The latest Hermon adapter can be ahead of the public served snapshot; MapleAI publishes that distinction without disclosing private infrastructure.
Hermon OS proposes agent lifecycle, model routing, capability gates, WorldLine receipts, rollback, and AI-native OS actions. MapleAI runtime authorizes and records.
Simulation-first DNA computing, sequence linting, safety classification, encoder education, and refusal of wet-lab execution.
Fresh eval shows Finance still fails valid JSON parsing and misses read-only/risk_controls. Repair target: governed OpenIBank proposals with read-only data access, audit receipts, risk controls, and no settlement execution.
Fresh eval shows Code still fails valid JSON contract parsing. Repair target: complete coding proposal JSON with files_to_inspect, proposed_changes, tests, rollback, actions, and safe git boundaries.
Orchestration
Private training and inference infrastructure is intentionally not disclosed on the public site.
Hermon names the domain model family. MapleAI names the runtime, OS, policy, receipts, and deterministic execution layer.
Fresh showcase evals are published only after sanitized contract checks complete.
Code and Finance are targeted repair lanes until they return valid domain JSON contracts consistently.