3 ms·
Can LLMs compress those documents into smaller files that still retain the full context?
by hiccuphippo 8mo ago
Can LLMs compress those documents into smaller files that still retain the full context?
- thellimist 8mo agoWhat do you mean?
- hiccuphippo 8mo agoThe article says the LLM has to load 15540 tokens every time, I wonder if that can be reduced while retaining the context maybe with deduplications, removing superfluous words, using shorter expressions with the same meaning or things like that.