Category: File Formats
Definition
JSONL (JSON Lines) is a text format where each line contains a valid JSON object, ideal for storing structured training data and model outputs.
How It Works
Each line in a JSONL file is independently parseable JSON, allowing streaming processing of large datasets. Common schemas include prompt-completion pairs for language models.
The format supports appending new data without parsing the entire file, making it efficient for continuous data collection.
Why It Matters
JSONL has become the standard for fine-tuning language models. Its line-by-line structure enables processing datasets too large to fit in memory.
OpenAI, Anthropic, and other AI providers use JSONL for their fine-tuning APIs and dataset formats.
← Back to File Formats | All Terms