I'm doing research about openAI's models for my Bc. Thesis?

Community
1

Hello,

I have been looking for further information about the models themselves. I was able to find the basics, such as the cost for tokens, maximum length of token vectors, etc. However, I didn’t really find any information about how the models were trained specifically (I’m aware of the GPT shortcut meaning), or if there are any tokens from the context window limited for OpenAI API services, etc.

I am pursuing this information because I am interested in integrating LLM models with the NAO robot. Additionally, I am curious about the handling of time interference. Any insights you could provide on these topics would be greatly appreciated.

Thank you.

2

You can look at the blog posts covering the release of new models. There you will find links to ‘technical reports’ which can be a starting point for your research.

Since the models are mostly proprietary and not open sourced it will be a challenge to get deep into the details of how the models have been built.

You can also refer to the inofficiall status page to learn about inference times for the different models over time.

1 Like
3

Thank you for your time i will look into it :slight_smile: .