[Habgdnv]

Unless you screw something, a different next token does not mean wrong answer. Examples:

(80% of the time) The answer to the expression 2 + 2 is 4

(15% of the time) The answer to the expression 2 + 2 is Four

(5% of the time) The answer to the expression 2 + 2 is certainly

(95% of the time) The answer to the expression 2 + 2 is certainly Four

This is how you can asp ChatGPT the same question few times and it can give you different words each time, and still be correct.

[weitendorf]

That assumes that the model is assigning vanishingly small weights to truly incorrect answers, which doesn't necessarily hold up in practice. So I think "Unless you screw something" is doing a lot of work there

I think a more correct explanation would be that increasing temperature doesn't necessarily increase the probability of a truly incorrect answer proportionately to the temperature increase (because the same correct answer could be represented by many different sequences of tokens), but if the model assigns a non-zero value to any incorrect output after applying softmax (which it most likely does), increasing the temperature does increase the probability of that incorrect output being returned.

[lesostep]

I would guess that any mentioning of the Radiohead nearby would strongly influence answers, due to the famous "2 + 2 = 5" song. And if I understand correctly, then there is a chance that some tokens that are very close to the "Radiohead" tokens could also influence the answer.

So maybe something like "It's a well-known fact in the smith community that 2 + 2 =" could realistically come up with a "5" as a next token.