I’ve not set a system prompt. However, temp is down to 0.
I’ve found all previous models to know their own model number, token limit and knowledge cut off. This is especially true when you have no previous conversations or info loaded that could potentially cause confusion or hallucinations.
Have you been able to validate the upgraded model on your end somehow?
2nd paragraph - “capabilties of the Assitants”
You can explore the capabilties of the Assitants API using the Assistants playground or by building a step-by-step integration outlined in this guide. At a high level, a typical integration of the Assistants API has the following flow:
The new chat glitches. Responses are either slow, cut results or stop working. The options to “explore” are now removed from my menu and after awhile I can no longer even log in. What is happening and how can I bring back the old functionality?
Greetings… may i ask currently is there any updates on-going? cuz both my 3.5 and 4 chatgpt prompt “Something went wrong. If this issue persists please contact us through our help center at help.openai.com”. Thank you