TODO: Model or adapter name

Intended use and limitations

TODO: Task, research users, supported input distribution, excluded uses, known failures.

Model and training

TODO: Base model and immutable revision; tokenizer revision and chat template; full-model or adapter configuration; recipe path and source commit; training data and split provenance; teacher identity/permissions where relevant; seed, budget, hardware, dependency freeze and container identity; maintainer.

Evaluation

TODO: Frozen baseline, benchmark and scorer identities, sampling budget, repeated trials, uncertainty, failures, cost/latency, report link, and domain expert review. Distinguish development replay from evaluation of generated model outputs.

Permissions

TODO: Selected artifact license and compatibility with base model, data, teacher outputs, and benchmark terms. Add the appropriate Hub license metadata only after it has been selected. State publication authority and visibility.