Prompt shouldn't include Character limit

[moritztim]

Feature request

Apparently LLMs might have a general understanding of roughly how much 50 characters is but since there's already a max_tokens defined, I don't think it helps to set a character limit, since it's not gonna result in more exact results than just relying on max_tokens.

Why?

It's not that big of a deal, I just thought it's an, admittedly small, waste of prompt tokens for every commit.

Alternatives

No response

Additional context

https://help.openai.com/en/articles/4936856-what-are-tokens-and-how-to-count-them

Using ShellGPT to test the character count accuracy of GPT 4

for i in {1..10}; do sgpt --chat --model=gpt-4 --no-cache 'write a sentence with exactly 50 characters' | wc -m; done 
45
53
55
56
57
53
60
57
54
57