Skip to content
MirrorFirm

What licence questions should an evaluation team settle before using generated material?

Evaluation teams should clarify permitted uses for source packages, derived benchmarks, model training and redistributed extracts before incorporating generated material into their workflows.

**Does receiving a sample establish rights for ongoing evaluation?**

Do not infer that from access alone. Ask which uses the proposed licence covers, including internal testing, repeated automated runs and access by contractors operating the evaluation system.

**What if the team modifies the tasks?**

Clarify treatment of derived task packs, translated policy text, changed schemas and new evaluator annotations. Ask whether those derivatives can be retained and reused separately from the original package.

**What about training and publication?**

State whether the material will inform model training, preference optimisation or evaluation only. Separately ask about publishing benchmark examples, tool traces and extracts in reports. Sharing an aggregate result and distributing the underlying task records are different activities and should be discussed separately.

**Does public server access change those questions?**

Material accessible through a public MCP server still needs an understood usage basis. Confirm which content the licence addresses and whether downloaded exports follow the same conditions.

**What should the first request contain?**

Provide a concise usage map covering users, storage, modifications and intended distribution. Request written clarification of the applicable rights and limits rather than assuming particular contract terms.

Translate intended evaluation activities into explicit licence questions without assuming that access permits every downstream use.

Audience:
Evaluation teams
Content type:
FAQ
Buyer question:
Scope and getting started

Content release: r-a220eb6e099b243b

More on Policy change propagation