🔒All inputs are processed locally in your browser. Nothing is transmitted, stored, or logged. Inputs disappear when you close the tab.

OCG-Industries Showcase · Cluster ④ AI Agents and Autonomous Systems

Fine-tuning
Lineage

A fine-tuned model is a base model plus a set of datasets, and six months later nobody can list them. Enter the base model, the datasets used with a licence flag for each, and this page derives a lineage key and tells you whether every dataset in the run carries a declared licence. A run with an undeclared dataset is flagged, not averaged away.

Data: SPDX licence identifier convention · no external data

Inputs STEP 1
Leave blank if the base model was not fingerprinted.
One per line as dataset-id:licence, for example docs-2025:cc-by-4.0. Use unknown when the licence was never recorded.
--
--
Lineage Status
--
Fully declared or incomplete
Declared
--
Share of datasets with a declared licence, display only
Distinct Licences
--
Count of distinct declared licence strings
Dataset Ledger
Detail
Fine-tuned model--
Base model--
Base fingerprinted--
Dataset count--
Declared count--
Undeclared count--
Undeclared dataset ids--
Distinct licences--
Declared (bp)--
Training steps--
Lineage key--
Lineage status--

Everything here is self-declared by whoever fills the form. This page records and checks a declaration. It does not inspect a dataset, verify a licence, or attest that the declared data is what actually went into training. Set one licence to unknown and run again to watch the declared percentage drop and the status flip.

execution_hash: --
Policy Mandate v2.0 · what is this?