Hello,
I’m using long term memory for in my code where I need more context about the “score”.
This is the schema of the database which was automatically created when I included the path.
sqlite> .schema long_term_memories
CREATE TABLE long_term_memories (
id INTEGER PRIMARY KEY AUTOINCREMENT,
task_description TEXT,
metadata TEXT,
datetime TEXT,
score REAL
);
My questions are, how is this score being generated? What is it related to? On what criteria and who is generating it? [Is my agent self evaluating it’s task performance or any internal program is doing it]? is it anonymous telemetry[from the documentation]? How is my LLM using this scores? Can my long_term_memory store my task’s outputs also?
Also when I was using memory I noticed the creation of short_term_memory, long_term_memory as well as entity memory for my agent without specifically initiating them in my code. Is that how it’s supposed to work?
Any information would be helpful.
Thanks in advance!