Fine-Tuning Datasets

Fine-tuning datasets are collections of instruction, question-and-answer or conversation pairs used to train or adapt a model. Their value depends on quality and licence, not size. Before you buy, look for a description of how the examples were written or collected, the format and number of records, a clear licence, and a sample you can compare with your own task.

• Customer Support Conversation Dataset (500 multi-turn dialogues) - $10.00
  500 original synthetic multi-turn support dialogues across 15 scenario types, for training a support assistant.
• General Instruction-Following Dataset (750 pairs) - $12.00
  750 original instruction/response pairs across 7 task types, for fine-tuning and evaluating instruction-following.
• Customer Message Intent Classification Dataset (600 examples) - $1.60
  600 original synthetic messages, evenly labeled across 10 intent categories, for training a message-routing classifier.