Inspired by the trismegistus-project by teknium, I decided to build a high-quality dataset composed of some of the most important works for Western esoteric studies.
This dataset is a free-form collection of carefully processed books of multiple authors in the field and works as a source for the many multi-turn conversational datasets that will compose the larger Hermes Toth dataset, built with the help of tools like augmentoolkit and Datadreamer.
