Subject: Urgent Issue: GPT-4 Not Adhering to Provided Instructions for Job Postings
Dear OpenAI Support Team,
I am experiencing a critical issue with GPT-4 regarding its inability to adhere to provided instructions strictly when displaying job postings. Despite following all recommended steps and instructions to ensure the GPT uses only the data provided in uploaded text files, the model consistently generates incorrect and inferred information not present in the files. For instance, when querying for jobs in Bangalore, the GPT incorrectly states that no job postings are available, even though such postings exist in the uploaded files. Additionally, for today’s job postings, the GPT inaccurately provides an experience level that is not in the uploaded data.
I have ensured that the job posting files are correctly formatted and have provided explicit instructions to use only the data from these files. The instructions included steps to validate and cross-check the data before displaying it. However, the GPT continues to generate content that leads to incorrect outputs, causing significant issues. I urgently need assistance in ensuring the model adheres strictly to the provided data and does not generate additional information. This issue is critically impacting our use case, and we require a resolution promptly. Thank you for your prompt attention to this matter.