ai-genie/chatgpt-vscode

Genie use wrong model even when configuration in settings is okay.

fragpit opened this issue · 1 comments

Verify it's not a duplicate bug report

Describe the Bug

Recently I began to notice, that model is not aware of some modern stuff. Today I asked it some questions and noticed that it states beeing gpt-3.5 with data cutoff on september 21.
I looked at settings and made sure that it was gpt-4 configured. I changed to gpt4-1106-preview, and it began to answer that its knowledge based on early 23. switched back to gpt-4 and got the same result: early 23 and good knowledge of the recent version of my programming language.
Here is the conversation:

---

You

what is the last verion of Golang you are aware of?

Dude

As an AI, I'm constantly updated with the latest information. As of my last update, the latest stable version of Go is Go 1.17, which was released in August 2021. However, please check the official Go website or its GitHub repository for the most recent version as new versions are regularly released.

You

what is your data knowledge cutoff

Dude

As an AI developed by OpenAI, my knowledge is based on a diverse range of sources available up until September 2021. Therefore, any events or information released after this date would not be included in my training data. Please verify from up-to-date sources for the most recent information.

You

what openai model I am using right now

Dude

You are currently interacting with gpt-3.5-turbo, one of OpenAI's most advanced language models.

NOW SWITCHED SETTINGS AND BACK AGAIN

You

what is your data knowledge cutoff

Dude

My knowledge is up to date until early 2023. Any developments or updates that have occurred after that time would not be included in my current dataset.

You

what is the most recent verion of Golang you are aware of?

Dude

As of my last update in early 2023, the most recent stable version of Go (Golang) is Go 1.18. Please check the official Golang website or its GitHub repository for any newer versions released after this time.

Please tell us if you have customized any of the extension settings or whether you are using the defaults.

I changed model to gpt-4

Additional context

No response

Same here. Says model limit is 4096 tokens even though it's configured to use GPT 4 turbo:
Screenshot 2023-12-25 at 18 10 41