Generative AI

AI Assistant

Developed a chat assistant with 4 different models namesly gpt 3.4 turbo, gpt-4, llama-2 and mistral. The chat assistant remembers the previous conversation and data provided to it. Taking into consideration the previous conversation it gives the appropriate answer.

Designed and developed personal chat assistant

Appropriate model is selected from the side panel. This selected model will work as a default LLM for the entire conversation
There we can see that the response of the selected LLM is formal. This is achieved by appropriate prompting
At the end of the conversation we can see that the model actually remembered my name and suggested exactly two surnames. This shows the LLM maintains the converstaion.

Chat with documents

Assume that we have 1000s of documents and we are searching for an appropriate information for a particular document. Now we need to check each and every document or the respective whole document. We make this task simpler here. We will upload all the documwnts in this chat with documnts and just ask the required question. It will fetch the exact answer from the entire documents.

Ease to read documents

Here we upload a set of documents. These documents are embeded and stored in a vector db.
Now we just need to ask any thing related to the dcuments and the assistant will fetch the appropriate details from the uploaded documents.
We can add upload any file like pdf, word file and excel. The fetched data can be muffled with any LLM model and can generate the responsive text to keep the converstion ongoing.