Open AmalLuke opened 4 months ago
SAME, I've tried 128k token windows ,it throw 'Unknown Error' , only working after readjust to 2048 tokens. MODEL: gpt-3.5-turbo; gpt-4-turbo-preview
I came to realize, that the max output token is 4096 tokens for most LLMs, input Token is larger, so maybe that is what is causing the issue.
Plugin works fine when Max Token is below 4000, but gives error when it is higher than that, even when using gpt-4-turbo-preview.