Alchemist provides a flexible and collaborative environment for dataset curation. Within an organization, any samples
can be used in a dataset, regardless of their association with specific projects. This approach maximizes the potential
for creating diverse and comprehensive datasets.
Collaboration is at the heart of Alchemist’s dataset curation process. You and your team members can work together on
curating datasets, collectively deciding which samples to include. This collaborative approach ensures that datasets
benefit from multiple perspectives and expertise.
As you work on a dataset, Alchemist provides real-time metrics to help you track your progress. These metrics include
the number of samples in the dataset, giving you a clear overview of your dataset’s size and composition. This
information helps you make informed decisions about when your dataset is sufficiently representative and ready for use
in instruction fine-tuning.