NLP inside your database - Query OpenAI’s GPT-3 from directly inside your database

@mikael_nida Thank you Mikael, glad you like it!

@mikael_nida thank you!

@mikael_nida Thank you!

@mikael_nida thank you!

@mikael_nida thanks

@pulkit_khanna Thank you

@pulkit_khanna thank you

@pulkit_khanna Thank you!

@pulkit_khanna great to hear! would love to see you in our community channels on slack!

@pulkit_khanna We are glad that you found MindsDB really helpful.

@nusimow Thank you

@nusimow thank you!

@nusimow always welcome in our community https://mindsdb.com/joincommunity

@nusimow Thank you!

@nusimow thanks

@camen Thank you. Feel free to join our https://mindsdb.com/joincommunity.

@camen Thank you!

@camen Thank you!

@camen thanks

@camen Thank you.

@stedmanblake One thing that can certainly help is the 'temperature' parameter. This is enabled with MindsDB - you can set it low to make sure the answer is at least more deterministic https://platform.openai.com/docs...

@stedmanblake Indeed. Solving these hallucinations is an active research area. In theory, architectures like DeepMind's RETRO should be better at this. MindsDB's approach could prove very useful here by providing a simple and fast integration with knowledge bases via our DB handlers.

@stedmanblake this is a very good point, I think it would be very cool to measure, certainly we should be able to ask the model if the info is made up or not, i will make some research on this, would you like to collaborate?

@stedmanblake MindsDB takes the accuracy and reliability of its predictions very seriously and implements a number of validation techniques to ensure that the output is trustworthy. Some of the validation techniques used by MindsDB include: 1. Data quality checks: MindsDB performs data quality checks on the input data to ensure that it is clean and suitable for model training. This helps to prevent issues with the output that can arise from poor-quality data. 2. Model performance evaluation: MindsDB evaluates the performance of its models on a validation dataset to ensure that they are performing well and not overfitting to the training data. This helps to prevent issues with the output that can arise from models that are too complex or not well-suited to the data. 3. Model interpretability: MindsDB provides interpretability features, such as feature importance, that allow you to understand how the model is making its predictions. This helps to ensure that the output is reasonable and that the model is not relying on obscure or unusual relationships in the data. 4. Post-prediction analysis: MindsDB provides tools for post-prediction analysis that allow you to examine the output of your models and make sure that it is accurate and makes sense. This helps to identify any issues with the output and ensure that the model is performing as expected. Overall, MindsDB takes a comprehensive approach to ensure the accuracy and reliability of its predictions and provides multiple tools and features to help you validate the output and ensure that it meets your needs.

@ankit_flamme Thank you

@ankit_flamme Thank you!

@ankit_flamme thanks!

@ankit_flamme Glad that you loved the product.

@andrii_osce Thank you Andrii; our team can help you where needed. Thanks for your support.

@andrii_osce If you need any help, you can join our community and team will help you https://mindsdb.com/joincommunity

@andrii_osce Thank you! Sounds great!

@andrii_osce sounds good. thanks

@andrii_osce Glad that you loved the product.

@afchavezt Thanks Felipe!

@afchavezt You can create a free demo MindsDB account (https://cloud.mindsdb.com/) to try it out yourself!

@afchavezt Thank you!

@afchavezt thanks!

@afchavezt Thank you. Feel free to join the slack community at mindsdbcommunity.slack.com

@timi_lab Thanks Tim.

@adam_carrigan You're welcome Adam

@timi_lab Thank you!

@timi_lab thanks!

@timi_lab Thank you.

@steevep Thank you.

@adam_carrigan @steevep thank you

@steevep We would love to hear your feedback once you've tried it. Thanks for all the support.