each llm interprets the prompt differently. can you please make a video in which we can have different prompt for gpt4o mini and separate system prompt if the voice agent fallbacks to llama3 for the llm? and also what are your thoughts on llama 3.3 for voice agents? how does it compare to gpt 4o when it comes to adhering to system prompt and function calling reliability?
![I.N "HALLUCINATION" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/n5B5q1Hwt_U/mqdefault.jpg)
each llm interprets the prompt differently. can you please make a video in which we can have different prompt for gpt4o mini and separate system prompt if the voice agent fallbacks to llama3 for the llm?
and also what are your thoughts on llama 3.3 for voice agents? how does it compare to gpt 4o when it comes to adhering to system prompt and function calling reliability?
guys with volume master at 600 it s perfect
You could have deleted the open ai key from open ai account in the middle of the conversation to test the fallback
That's actually right. Haven't thought much while recording it. But that's a good thought. Thanks
Your voice is very low in general and especially compared to the LLM voice.
Right
eat breakfast before you record so u hv energy to speak.
and you fake mic is flowing through your body so better buy 1 mic