🤯 Using Wolfram Alpha Plugin to Train/Finetune Custom OpenAI Models in one shot

This is crazy. I can now fully train models using Wolfram, in this example in a single prompt.

Using the Wolfram Alpha plugin, perform a scientific calculation (e.g., calculate the speed of light in a medium), then use the OpenAI API to fine-tune a language model for scientific explanations, and generate an explanation for the calculation result. Scientific calculation, Max tokens for explanation

8 Likes

You’re a machine, brother. Remarkable stuff.

I think this might be my best plugin yet. It does some crazy stuff.

Are you sure that isn’t just a hallucination? Also, if not, what model did it pick, out of curiosity, and can you use it in the playground?

These are actual requests made to the OpenAi api, you can control all settings to eliminate hallucinations. Give it a spin and let me know what you think.

This is really awesome and valuable.
Thank you for sharing.