#ChatGPT terrible at writing code with the openai api

1 messages · Page 1 of 1 (latest)

crystal elbow
#

I assume that this is because of the knowledge date cutoff. Has anyone found good workarounds? uploading socumentation?

sharp oracleBOT
#

Hi @crystal elbow. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:

#

If you're finding that ChatGPT isn't performing as well as you'd like when it comes to writing code with the OpenAI API, there are a few strategies you might consider:

  1. Improve your prompt instructions: Make your instructions clearer, more specific, and provide examples if possible. This can help guide the model to generate the desired output.

  2. Use a different model: If you're currently using a smaller model like gpt-3.5-turbo, you might consider switching to a larger model like GPT-4 if your use case requires a larger context window.

  3. Fine-tuning: OpenAI provides the ability to fine-tune models, which could help improve performance on specific tasks. You can review the fine-tuning guide (https://platform.openai.com/docs/guides/fine-tuning) for more details.

  4. Upload Documentation: If you're using a GPT that supports knowledge files, you could consider uploading relevant documentation as a knowledge file. However, be aware that there are limitations to the number and size of files that can be uploaded, and some users have reported issues with GPT's ability to effectively use large amounts of text from knowledge files.

Remember, GPT models are not perfect and their performance can vary depending on the specific task and how they're used. If you're encountering issues, it might be worth reaching out to OpenAI for support or checking the OpenAI Developer Forum (https://community.openai.com/) for advice from other users.