we are going to use chatgpt for abstarcting legal document from chatgpt and need to create a template but chatgpt is a text based model
so it can only respont in text response way.
i need information from chatgpt about the legal document rules of an state but it is responting with
examples :
As an AI language model, I don’t have direct access to my training data or know where it came from. However, I was trained on a mixture of licensed data, data created by human trainers, and publicly available data. OpenAI, the organization behind my development, has not publicly disclosed the specifics of the individual datasets used, including whether data from legal land documents in state was used.
is their any other way to train a chatgpt recording legal document rules
Cr BhargaviPosted Jul 22, 2023, 2:25 AM
With just Chatgpt to be trained as expectation I am not sure if we will be able to achieve it. You can start using the services which are available in OpenAi, Azure openai etc and get those things trained. When it comes for Azure OpenAi services, it's so easy to solve the things and train your model such that you just need a subscription and knowledge of Azure and OpenAi services. You can go ahead and try building models of yourself aswell like you can use NLP and with the help of the NLP you would be able to have the LLM models also in place and implement the solution and build by yourself.
Chris LovePosted Jul 18, 2023, 3:10 PM
So, you won't really be able to train it in the sense you can build a customized model. What you can do, if you have a pro account is use the GPT4 model with the code interpreter. You can then upload your document(s) to analyze and have a conversation about that document or data.
If you want to train a custom GPT-4 model you can do that through the OpenAI API. But I have to be honest, I think it is super expensive.
There are other options to build a private model on a LLM that would be more cost-effective IMO.
My AI team is working on some content around using the code interpreter to analyze data. We also have some content in the pipeline on using Hugging Face for your own models. That might help, so watch for those on the home page of C# Corner soon.