I am currently working on building a chatbot using OpenAI, and I would appreciate some guidance on an issue I’m facing. The goal of my chatbot is to simulate conversations with a well-known person,to make the chatbot more interesting and engaging, I decided to fine-tune the model by providing it with additional information using a custom dataset. However, even after this fine-tuning process, the chatbot is still not responding in the manner I anticipated.
Are there any specific techniques or approaches I should consider? Has anyone else encountered a similar issue while building a chatbot?
Any advice, insights, or experiences you can share would be greatly appreciated. Thank you in advance for your help!
Without details about what you’ve already done, specifics about your fine-tuning, what you expect the model to do, and a description of how it is failing to meet your expectations no one is going to reliably be able to give you any advice to improve your results.
I remember asking this very same question over 6 months ago. Maybe the better approach is embeddings, with all the content you have on this person and a prompt which specifies that responses can only be made in the persona of the person.
