[Feature Request] Message/Prompt compression (not using LLM)

[Wendong-Fan]

Required prerequisites

• [X] I have searched the Issue Tracker and Discussions that this hasn't already been reported. (+1 or comment there if it has.)
• [ ] Consider asking first in a Discussion.

Motivation

for a large volume of message/prompt information instead of using LLM do the summarize there's also other solutions like https://pypi.org/project/gptrim/ , we need to do study and pick a better solution

Solution

https://arxiv.org/abs/2210.03162 https://pypi.org/project/gptrim/

Alternatives

No response

Additional context

No response