No post body was provided.
Untitled topic
1–5 of 5 posts
Re: undefined
#2I don't understand what this chat transcript is demonstrating.
Re: undefined
#3Never expect factual accuracy from large language models. No one actually knows how to make neural networks which always conform to a well-defined knowledge graph and so that means you should always assume all of the output contains subtle errors that you won't notice on first glance.
Re: undefined
#4Never expect factual accuracy from large language models. No one actually knows how to make neural networks which always conform to a well-defined knowledge graph and so that means you should always assume all of the output contains subtle errors that you won't notice on first glance.
This is eternally the correct response.
Re: undefined
#5Interacting with LLMs feels like two people talking past each other. OpenAI is so scared of negative press the end product is weird to use. Would love to use the models with fewer handcuffs on them.