This is the small version, suitable for tiny language models. For use with bigger models, take a look at full Alpaca Cthulhu data set.
Answers in the original data set was rewritten, as if the answer was given by a dedicated, high ranking cultist in the Cult of Cthulhu. The instructions were to keep the replies factually the same, but to spice things up with references to the Cthulhu Mythos in its various forms.
This was done as part of a learning experience, but I am making the data set… See the full description on the dataset page:
https://huggingface.co/datasets/theprint/alpaca_cthulhu_small.
