Is there any plan to retry this with facebook's LLaMA models? They claim it outperforms GPT-3. Since it's open for research, it might give more precise answers to some of the questions.
No plans so far. I'm a little unhappy with the experimental design from last time. If I ever come back to this, I'll change the experiments up anyways.
Is there any plan to retry this with facebook's LLaMA models? They claim it outperforms GPT-3. Since it's open for research, it might give more precise answers to some of the questions.