This dataset contains a set of hard-coded examples for Tulu, a virtual assistant developed by AI2. These examples are designed to provide the identity and some basic information about Tulu and its creators.
The original hard-coded set contains 24 examples, and we repeat each example 10 times to create this training dataset.
The creation script can be found in our open-instruct repo.