JevMade Sign in
← Back to guides

JevMade field notes / Architecture and workflow guide

Deciding when an AI assistant should clear its working memory

This guide explains how a plugin helps an AI assistant decide when to clear its working memory. It combines two simple questions to suggest the safest time to save space without losing important details.

Original by kunchenguidEvaluation

Listen to this guide

JevMade’s plain-English explanation

0:00 /

Our summary

This guide explores a plugin that tells an AI assistant when to clear its working memory. People use it to find natural stopping points in a conversation. This prevents the software from running out of room or deleting important details while a task is still active.

The plugin uses Jev, an AI tool that chooses from options rather than writing an answer. It asks if the current task is finished and what kind of work is happening. The software combines these answers into a score to decide if clearing memory is a safe choice.

The required score changes based on how full the memory is, becoming less strict as space runs out. Custom instructions always start by only suggesting action until a person checks them. This setup is useful for people managing long sessions with an AI assistant across different programs.

Key takeaways

  1. Combine questions about task progress and work type instead of asking one vague question about clearing memory.
  2. Require higher certainty to clear memory when space is mostly empty, and relax it as space fills up.
  3. Keep custom instructions in a suggestion mode until you test them to prevent unwanted automatic actions.

The repository authors reported the behavior and measurements from their own tests.

Repository README and documentation

Read the original guide Opens the author’s site in a new tab.

Keep this for later

Sign in to bookmark experiments, guides and videos, and keep notes only you can see.

Continue to sign in

We’ll bring you back to this listing.