Large language models

I really don’t get all that hype around large language models. I can understand their uses in NLP (and their superb transfer learning powers), but people are treating them like the end-goal of AI. I might be biased, but they don’t look that smart to me – I’ve tried this one here: https://6b.eleuther.ai/. You get different results based on what parameters you use, but:

Model with top-p = 1.0, temperature = 0 (i.e. choose best – I guess that’s what it means – you get really bizarre results with the default website parameters)

Q: Is carbon needed for steel?

A:I have been reading a lot about steel lately and I have been wondering if carbon is needed for steel. I have read that carbon is needed for steel to be strong and hard. I have also read that carbon is needed for steel to be ductile. I have also read that carbon is needed for steel to be magnetic. I have also read that carbon is needed for steel to be rust resistant. I have also read that carbon is needed for steel to be heat resistant. I have also read that carbon is needed for steel to be corrosion resistant. I have also read that carbon is needed for steel to be weldable.

Q: How do I make a hamburger?

A: You can use the following code to make a hamburger: <div class=“hamburger”> <span></span> <span></span> <span></span> </div>

.hamburger { display: inline-block; width: 40px; height: 40px; margin: 0; padding: 0; border-radius: 50%; background-color: #