What should a buyer ask for in a first sample of facility material?
A checklist of eight written requests for a first sample, covering cases, tables, reference trajectories, graders, labels, reset, generation method and licence wording.
Put these requests in writing when you ask for a sample, so the first review tests what you actually need.
- **Case file.** Prompts, policy versions and expected outcomes for each case, including cases where booking is the right answer. - **Start-state tables.** Your preferred format, with a schema file listing types, keys and nullability. - **Reference trajectories.** Expert-style steps, with the role taking each action. - **The grader.** A description of what each check covers and what it ignores. - **Exception labels.** Tags so you can slice results by defect type. - **Reset procedure.** One you can run yourself, with a note on how seeds work. - **Generation statement.** How the material was generated and what disclosure review was applied. - **Licence wording.** The proposed text for your intended uses, such as training and published evaluation.
Putting these requests in writing means the first review tests what you actually need instead of what happens to be easiest to hand over.
- Audience:
- Model labs
- Content type:
- Checklist
- Buyer question:
- Scope and getting started
Content release: r-a220eb6e099b243b
Related questions
- What does a first pilot look like for an agent team in vendor finance?
- What limits apply, and will we receive any real customer records?
- How could a pilot be delivered, versioned and wired into our CI?
- How should a team scope and license a generated booking pack?
- What should we ask for in a sample or pilot of generated booking material?
- How should a sample be delivered, versioned and load-checked in my pipeline?