Repeated error and confusion in responses - Request for clarification to improve the model

ChatGPT
1

Hello,

I would like to submit feedback regarding a repeated error I encountered during my session with the AI (ChatGPT). The error relates to confusion about the processor model Intel Core Ultra 9 14900K and its compatibility with my MSI MEG Z890 GODLIKE motherboard, as well as the fact that this error persisted despite the AI assuring that the context would be managed properly within the same session.

Problem Details:

  1. Initial Request: I asked several questions regarding the best processor choice for an MSI MEG Z890 GODLIKE motherboard, intended for 4K gaming and streaming. The socket of my motherboard is LGA 1851, and I was looking for a compatible Intel Core Ultra 14th generation processor for this setup.
  2. The Repeated Error:
  • Several times, the AI mentioned an Intel Core Ultra 9 14900K, which is incompatible with the LGA 1851 socket, as it uses the LGA 1700 socket. This model does not exist in the Intel Core Ultra range (14th generation), which led to confusion.
  • Even after I pointed out this mistake, the AI continued to mention the 14900K in its responses.
  1. Context Management Issue within the Same Session:
  • When I first pointed out the error, the AI assured me that errors would not be repeated within the same session, as it was supposed to keep track of our conversation history to avoid repeating mistakes.
  • However, despite this assurance, the AI kept mentioning incorrect models (such as the 14900K) in its subsequent answers, showing inconsistency in applying this context management rule.
  1. Impact of the Error:
  • This confusion caused me uncertainty and slowed down my decision-making, as the AI was not providing accurate information about processors compatible with my LGA 1851 motherboard.
  • Despite my patience and clarifications, this error persisted, even after several attempts to get the correct information.

Conclusion and Suggestion:

I understand that errors can happen, but it is important that the AI correctly manages context within a single session to avoid repeating previous mistakes. It seems there was an issue with context handling and continuity in this session.

I believe an improvement in the context management of the model could prevent such confusion and provide more accurate and relevant answers for complex requests like this one.

Thank you for considering this feedback to improve user experience in the future.