Cheapest GPT-4 Turbo, GPT 4 Vision, ChatGPT OpenAI AI API
Door
NextAPI |
Bijgewerkt 23 дня назад |
Artificial Intelligence/Machine Learning
Populariteit
9.8 / 10
Latency
195ms
Serviceniveau
100%
Health Check
100%
Terug naar alle discussies
Finish reason: length
paywithcode
месяц назад
Greetings!
i am working with this paramethers:
"role"=>"user",
"model"=>"gpt-4-turbo-2024-04-09",
"max_tokens"=> 100,
"temperature"=>0.9
but i getting answer that contains 370-400 chars and finish reason - length.
What should i change to get a bigger answer?
Doe mee aan de discussie - voeg hieronder een opmerking toe
Log in / Schrij u in om nieuwe opmerkingen te plaatsen
Hey there,
Glad you figured out the answer. Yes that is correct, the “max_tokens” object is similar to words, it limits how many words GPT-4 can reply with. If you increase it you can use it to generate longer answers.
If you do have any other questions feel free to message us!
Best regards,
The NextAPI Team
max_tokens 500 is better. sorry for disturbing)