No, all you need to do is memorize enough vocabulary and have enough chances to get lucky on occasion. You can have better odds if you learn the conditional probability of that vocabulary given surrounding words.
And this is all that GPT is doing.
And this is all that GPT is doing.
To do it it, you need a form of compression: reasoning, intelligence, knowledge.
We already know LLMs tokenize words, so they have definitely generated and stored such a table. The marvelous step forward from previous word-based solutions is that they store not just single words, not just pairs of words, but larger sequences of words, even up to short sentence length. It makes the table pretty large, as you suggest, but that's exactly we've been seeing, and each generation of GPT is that much larger than the one before.