Mar 27, 2024 7:53 pm
We have a version of "Neo" that's now trained on a base model which was trained on the entire library of PubMed research articles. We are not yet done augmenting this model and have only completed one epoch of training so far. However, even at this level, the response is quite detailed.