me3meme

3 ms·

1.
by me3meme 1y ago
Metaphor: finding a path from a initial point to a destination in a graph. As the number of parameters increases one can expect the LLM to be able to remember how to go from one place to another and in the end it should be able to find a l
2.
Ask HN: Books for LLMs
1 points by me3meme 1y ago|0 comments
3.
by me3meme 1y ago
if people are going to read 630 comments about this post, I think they are not willing to read the prompt they prefer to expand endlessly like a LLM.
4.
by me3meme 1y ago
-- Do a 2x2 block matrix multiplication to convince yourself that this preserves the form achieved so far. I don't consider this a proof. Perhaps you have in mind two simple but key properties of reflections about the hyperplane ortho
5.
by me3meme 1y ago
I just selected lecture 07 to take a look: Lecture 07 is about QR factorizacion and Householder reflections. The author proves how to construct a reflection to make zeros in the first column and then he just claims that following this proc
6.
by me3meme 1y ago
I think we must use conditional probability, that is the integral of p(X|A) P(A), for example probability the prior event was 5 minutes ago probabity(the next one is 10 minutes from the previous one (that is 1/2). This is like markov
7.
by me3meme 1y ago
From last event to this event = 10, from this event to next event = 10, so the time between the first and the third event is 20, where is the surprise in the Waiting Time Paradox?, sure I must be missing some key ingredient here.