§ Training Data · Ground Truth Aggregator
Ground Truth Dashboard
Review and export approved model outputs as ground truth training data for fine-tuning.
Total Samples
0
Approved
0
Rejected
0
In Golden Set
0
Models
0
Filters:
No samples match filters
Approve EvalSamples to see them here
📊 Export format: JSONL (JSON Lines) — one record per line. Use for fine-tuning datasets or custom training pipelines. Each record includes prompt, response, expected output, and review metadata.