[hiccuphippo]

Can LLMs compress those documents into smaller files that still retain the full context?

[thellimist]

What do you mean?

[hiccuphippo]

The 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.