search

Messerschmidts101 / Call_Middleman

A middleman solution for call centers that facilitates communication between agents and customers. This tool filters, summarizes, and enhances messages, helping to reduce cognitive load and emotional stress for call center agents. By streamlining interactions, it improves both the mental well-being of agents and their overall work effectiveness.
0 stars 0 forks source link

Long Memory Chat History #4

Closed Messerschmidts101 closed 1 month ago

Messerschmidts101 commented 2 months ago

As the chat history shares input tokens with the LLM, it’s important to manage the token limit efficiently. Instead of including the entire conversation history, we need a mechanism to summarize or extract only the most relevant parts of the chat. This will ensure that the LLM maintains context without exceeding token limits or including unnecessary information. Implementing a chat history summarization method is essential for optimizing token usage and improving model performance.

Assigned to: @Messerschmidts101

Messerschmidts101 commented 1 month ago

This issue was resolved in commit [e69b6a63e77d88be4c1509edd2bd806f661cc69d]. Closing this issue now.