File size: 3,721 Bytes
5a19b22 97210a7 5a19b22 97210a7 5a19b22 97210a7 5a19b22 97210a7 5a19b22 97210a7 1e25453 5a19b22 97210a7 5a19b22 97210a7 5a19b22 97210a7 5a19b22 97210a7 5a19b22 97210a7 5a19b22 97210a7 5a19b22 97210a7 5a19b22 6bbe2f1 5a19b22 97210a7 5a19b22 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 | ---
pretty_name: Brief 003 - Marketing brief review
language:
- en
tags:
- marketing
- synthetic
- human-review
size_categories:
- n<1K
configs:
- config_name: default
data_files:
- split: train
path: data/briefs.csv
usecols:
- id
- user_text
- deliverable
- product_description
- offer
- audience_description
- awareness
- evidence
- prohibited
- status
- industry
- price_point
- completeness
- approval
dataset_info:
features:
- name: id
dtype: string
- name: user_text
dtype: string
- name: deliverable
dtype: string
- name: product_description
dtype: string
- name: offer
dtype: string
- name: audience_description
dtype: string
- name: awareness
dtype: string
- name: evidence
dtype: string
- name: prohibited
dtype: string
- name: status
dtype: string
- name: industry
dtype: string
- name: price_point
dtype: string
- name: completeness
dtype: string
- name: approval
dtype: bool
---
# Brief 003 · Marketing brief review
Review the generated marketing requests and record approval directly in the dataset. This batch contains 339 briefs. All products, offers, and supporting claims are fictional.
## Start reviewing
**[Open the brief viewer and editor →](https://huggingface.co/datasets/zeton-labs/marketing-briefs/viewer/default/train)**
1. Click a row to inspect the brief. Start with **user_text**, the request a user would send to the model.
2. Compare it with the structured brief: **deliverable**, **product_description**, **offer**, **audience_description**, **awareness**, **evidence**, and **prohibited**.
3. Click **Toggle Edit Mode** in Data Studio. Set **approval** to `true` to approve or `false` to reject. Leave it `null` until reviewed. Editing requires a Hugging Face account with write access to this repository.
4. Click **Commit** to save your edits. Until committed, changes are only staged in your browser.
| Approval | When to use it |
| --- | --- |
| `null` (blank in CSV) | Not reviewed yet, or still needs a decision. |
| `true` | Coherent, faithful to the structured brief, and useful as a training example. |
| `false` | Contains contradictions, invented details, missing constraints, or another quality problem. |
**approval** is the review decision. The older **status** column is retained for compatibility and does not need to be updated.
## What to check
- Does **user_text** sound like a believable request and ask for the correct deliverable?
- Are product facts, pricing, audience, evidence, and prohibitions preserved without invented details?
- Does the awareness stage match the information provided? A blank awareness field is expected when the request does not name a stage.
- If **completeness** is `needs_clarification`, does the request preserve the unresolved choices instead of inventing answers? Missing information is intentional in some examples.
## Tips
Use search to find a brand, deliverable, or phrase. Click a row and copy its URL to share that specific brief. The split is named **train**; this is a review queue and approval is still required before training.
Keep **id** and the column names unchanged. Only rows with **approval** set to `true` should be considered approved for training. Edits committed here update `data/briefs.csv` on Hugging Face; they do not automatically update the local CSV or JSONL. Republishing with `--force` replaces the remote CSV, including review edits.
[Hugging Face editing instructions](https://huggingface.co/docs/hub/en/datasets-cell-editing)
|