Tool
JSONL field viewer and row sampler
Top-level keys are counted. The first n lines, or n lines from a seed hash, are kept. This page does not validate or convert formats. Lines never leave this tab.
Sampling runs in this tab. GearDel does not receive the JSONL.
How a sample is chosen
Each non-blank line must be a JSON object. Top-level keys are counted across parsed rows. Nested keys are not flattened.
Head mode keeps the first n parsed objects in paste order. Seed mode hashes seed + NUL + JSON.stringify with FNV-1a (the same hash as the split page), sorts, and keeps n rows, then shows them in original line order.
The prompt column is Alpaca instruction plus input, or the first user/human turn. Other shapes leave that column blank.
Blank lines are counted and skipped. A leading JSON array is rejected. At most 20,000 lines or about two million characters.
What this page does not do
- It does not convert Alpaca, ShareGPT, or OpenAI messages. Use the JSONL checker for that.
- It does not stratified-sample by a label field. The seed hash is not a statistical draw of the whole corpus.
- Field examples are clipped previews, not a schema inferred from every nested object.