About setting the " /tokens [n] Set/Display tokens to use." parameter

[marioseixas]

It is possible to not set this parameter so it could default to Inf, Inf meaning (4096 - prompt tokens)? That way no one needs to keep making calculations about how many tokens to use or to make available:

[GaryDean]

Set max_tokens=-1 (/tokens -1) for maximum allowable response tokens.

When /tokens is -1, max_tokens is calculated thus:

token_limit = 4096   # for gpt-davinci dan gpt-3.5-turbo
used_tokens = calc_total_tokens(system_prompt + user_conversation))
max_tokens = token_limit - used_tokens

Does this address your question?