Independently recomputes a model's reported outputs from its declared specification (transform, intercept, coefficients) and stated inputs, then diffs the recomputation against the model's own reported outputs within your tolerance. Returns replicated or not-replicated, with the failing segment named. If the declared specification can't be recomputed as stated, the honest result is not-replicable-as-specified — not a forced pass or fail. Step 1 of the Model Validation Replication Cycle, where anything other than a successful replication puts the cycle on a §27 hold gate.
Copy this paragraph into Claude, OpenClaw, or any MCP-aware agent to run this exact tool, with this sample, and verify the artifact.
Run the AINumbers MCP tool `replicate_model_outputs`. Task: Independently recomputes a model's reported outputs from a declared model specification (version, as-of date, transform, intercept, coefficients) and a stated input record set, then diffs the recomputation against the model's own reported outputs within a caller-supplied tolerance.
Call it with arguments: {"policy_parameters":{"model_spec":{"version":"credit-scorecard-v3.2","as_of_date":"2026-01-01","transform":"linear","intercept":10,"coefficients":{"x1":2,"x2":-1}},"tolerance":{"abs_tolerance":0.5,"rel_tolerance":null},"records":[{"id":"r1","segment":"prime","features":{"x1":5,"x2":3},"reported_value":17},{"id":"r2","segment":"prime","features":{"x1":2,"x2":4},"reported_value":9.8}]}}
Verify before trusting: call `verify_execution_hash` on mcp.ainumbers.co (https://mcp.ainumbers.co/mcp) with the parameter `claimed_hash` set to the returned `execution_hash`, passing the full artifact the run returned (the object containing `policy_parameters` + `output_payload` + `execution_hash`; equivalently `policy_parameters` + `output_payload` with `claimed_hash`), not the bare hash string.
Return the ledger link https://ledger.ainumbers.co/ so a human can re-verify without contacting us.
PII rule: All inputs are processed locally in your browser. No data is transmitted. Do not enter real personal data — use synthetic or anonymised inputs only.
Open the tool with the sample prefilled: https://ainumbers.co/chaingraph/art-488-model-replication-diff.html#p=v1.H4sIAAAAAAAA_wFuAZH-eyJtb2RlbF9zcGVjIjp7InZlcnNpb24iOiJjcmVkaXQtc2NvcmVjYXJkLXYzLjIiLCJhc19vZl9kYXRlIjoiMjAyNi0wMS0wMSIsInRyYW5zZm9ybSI6ImxpbmVhciIsImludGVyY2VwdCI6MTAsImNvZWZmaWNpZW50cyI6eyJ4MSI6MiwieDIiOi0xfX0sInRvbGVyYW5jZSI6eyJhYnNfdG9sZXJhbmNlIjowLjUsInJlbF90b2xlcmFuY2UiOm51bGx9LCJyZWNvcmRzIjpbeyJpZCI6InIxIiwic2VnbWVudCI6InByaW1lIiwiZmVhdHVyZXMiOnsieDEiOjUsIngyIjozfSwicmVwb3J0ZWRfdmFsdWUiOjE3fSx7ImlkIjoicjIiLCJzZWdtZW50IjoicHJpbWUiLCJmZWF0dXJlcyI6eyJ4MSI6MiwieDIiOjR9LCJyZXBvcnRlZF92YWx1ZSI6OS44fV19Dbx_8G4BAAA