|
|
|
|
|
|
by aarong11
1047 days ago
|
|
|
According to several sources, GPT-4 is already doing this (It's called "Mixture of experts"). I agree though, it's still useful on an application level. You can build out even higher on that and have the LLM decide which is the most appropriate "hats" to wear for a given task. |
|
|