Whilst we’re all context wrangling, turns out that there’s not a massive amount of evidence that larger context sizes will even allow LLMs to deal with all the information sensibly. pic.twitter.com/iLMrb1g9N8
— Replit ⠕ (@Replit) 17 juillet 2023
Whilst we’re all context wrangling, turns out that there’s not a massive amount of evidence that larger context sizes will even allow LLMs to deal with all the information sensibly.