How did you manage to get such a low score without consciously trying to make it bad…? It’s like if you got to race a car and took one hour longer than everyone else and then bragged about it like it’s good.
It is, it’s using a llama model in the API. If you watch the network tab you can see the calls it makes and modify them yourself, they’re simple get requests. It’s kinda fun to pick random words and see what the LLM comes up with.
neal.fun
Ważne