[visarga]

It's useless before the model gets instruction and preference tunings. Won't even follow a simple ask, it will just assume it is a list of questions and generate more, or continue with slightly related comments.

FB trained a LLaMA-I (instruction tuned) variant for sports, just to show they can, but I don't think it got released.

[cfcf14]

You have to prompt it correctly, non-instruction-aligned models don't behave like agent simulators by default.

[qingdao99]

Surely it would work with a format like:

User: <question or task>

Assistant:

[ShamelessC]

Useless!? C'mon.

[RugnirViking]

it's not that useless, you just have to prompt it the right way (usually by offering an example of the kind of output you want)

[notpushkin]

So, you need to know how to tune it.

[ma2rten]

It's still useful, but you need to know how to use it.

[skhm]

Try giving it a simple request instead ;)