I think about something like auto-compaction. Can we somehow add tokens count to the llm results?
I think about something like auto-compaction. Can we somehow add tokens count to the llm results?