Massive Update - LLMs Updated, Superior Quality and Detail, Negligible Error Rate

Today marks a significant change to Scribeberry's AI medical model. Over the past few months, we have been relentlessly speaking with you, our users, to measurably improve our product.

We are proud to announce that we have permanently trained and refined our AI models to allow for massive improvements. Below are three of the key improvements we would like to share with you today.

Hallucination Rates: 7% (Old) vs 0.05% (Now)

Now, our AI will be able to recognize its own limitations with almost 100% accuracy and either not input new information, or clearly warn you that the information is inferred based on the history provided.

Details: More Thorough Notes

In our previous updates, we expanded the limit of conversations from roughly an hour to 7.5 hours. In this update, we have improved the context recognition and detail recollection to ensure the final generation retains all relevant contextual detail, even in longer conversations or dictations. This will ensure quality no matter the quantity.

Source Data: Knowledge up to August 2023

We have updated our LLMs (large language models) source data with more recent and relevant references. With data updated up to August 2023, Scribeberry will be able to generate your notes while using the most up-to-date and relevant details about your cases.

Ultimately we hope the above changes reflect an improvement in your overall workflow and speed and help continue to save you time. As always, if you have any feedback or questions, please email us at any time.