[HN Gopher] Understanding Large Language Models - A Transformati...
       ___________________________________________________________________
        
       Understanding Large Language Models - A Transformative Reading List
        
       Author : mariuz
       Score  : 62 points
       Date   : 2023-02-11 17:34 UTC (5 hours ago)
        
 (HTM) web link (sebastianraschka.com)
 (TXT) w3m dump (sebastianraschka.com)
        
       | mensetmanusman wrote:
       | Did anyone predict the unreasonable effectiveness of LLMs?
       | 
       | We seemed to have stumbled upon something quite big as humans...
        
         | loveparade wrote:
         | I worked in one of the big labs when the first large models
         | came out and I can pretty confidently say that nobody in the
         | field predicted this. Sure, there were always people who said
         | "let's make models bigger because why not, we have the infra
         | and it'll be a good paper" but nobody expected them to become
         | _this_ good just by being bigger and using more data. The
         | consensus was that they 'd hit a ceiling of what they can do
         | much sooner.
        
           | tomr75 wrote:
           | aren't humans trained on massive data sets, why not expect
           | the same for digital intelligence?
        
             | soraki_soladead wrote:
             | Only some model architectures continue to get better as you
             | pump in more data. Transformers and their variants have
             | this property more so than prior architectures.
        
         | tough wrote:
         | Or maybe language just isn't that complex at all?
         | 
         | Same with pixels?
         | 
         | Dunno
        
         | GeorgeTirebiter wrote:
         | Perhaps Monica Anderson understood early? https://experimental-
         | epistemology.ai/the-red-pill-of-machine...
        
         | corbulo wrote:
         | We stumbled upon a mirror of our own cognitive bias.
         | 
         | Another chapter in the era of Narcissus and Echo.
        
       | auggierose wrote:
       | Yeah, I was looking for a book about this topic, but there
       | doesn't seem to be anything out there except for research
       | articles, it is just too new.
       | 
       | When I look at the descriptions of the papers, it sounds
       | incredibly complicated, and at the same time incredibly trivial.
       | Clearly this whole domain is not well enough understood yet to be
       | explained properly. Or maybe there is no interest yet in clear
       | and concise explanations.
        
         | CongoDan wrote:
         | That's how blockchain was a few years back. Now everyone and
         | their uncle has a "how blockchains work" article or video
        
           | corbulo wrote:
           | And yet I feel like I know less at this point than I did
           | then.
        
         | KHRZ wrote:
         | The description of the attention mechanism of GPT architectures
         | and a couple of examples can be very brief. Then you have to
         | supply your imagination and realize that the model simply found
         | a whole bunch of very effective such attention measures by
         | itself, which are all computed for every query, and they could
         | be anything from the straight forward examples, or some more
         | clever abstract types of attention. I think we'll need some
         | more descriptions of what the more important attention measures
         | it comes up with really are, to understand better.
        
       | qup wrote:
       | > Large language models have taken the public attention by storm
       | - no pun intended.
       | 
       | As far as I can tell, no pun made, either!
        
         | marginalia_nu wrote:
         | "Attention" is the would be pun, I think.
        
         | xwowsersx wrote:
         | The pun is in "attention" because GPT uses "attention" to weigh
         | each input token and comes up with an attention score between
         | whatever token is currently being generated and _all_ the input
         | tokens then it 'll take those scores to determine what weight
         | each contributes to the output. Something along those lines...
         | I'm no GPT expert.
        
         | 0z8 wrote:
         | To make a pun in this domain, attention is all you need!
        
       ___________________________________________________________________
       (page generated 2023-02-11 23:01 UTC)