Just how many LLM papers are being published? (**A LOT**)

286 views
Skip to first unread message

Alan Timm

unread,
Oct 5, 2023, 2:12:24 PM10/5/23
to RSSC-List
Up to this point I've been relying on a handful of YouTube personalities to clue me in on recent developments in the AI space.  It's nice.  They talk, I listen, and occasionally I download the paper to read it.  Or have adobe acrobat reader read it to me.  I may have to take a different strategy.

To satisfy my curiosity I thought I'd use the handy dandy search feature of arXiv to see how many LLM papers are being published.  This is a simple keyword search using the term "llm".  Here are some stats:

There have been 175 papers published THIS MONTH (the past 5 days):

There were 696 papers published LAST MONTH:

And to wrap it up, there have been 3005 total published papers so far THIS YEAR:

Jim DiNunzio

unread,
Oct 5, 2023, 8:59:27 PM10/5/23
to Alan Timm, RSSC-List
Sounds like you need that supposedly near infinite context LLM to feed in batches of LLM papers (due to memory limit?) to rank them based on your interests or other factors,  then sort by rank. Use an LLM to find the best LLMs... the snake is eating its own tail...

Jim

--
You received this message because you are subscribed to the Google Groups "RSSC-List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rssc-list+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/rssc-list/0f061a95-958a-4bd2-93f7-c5fd38462ccbn%40googlegroups.com.

Alan Timm

unread,
Oct 6, 2023, 1:07:07 AM10/6/23
to RSSC-List
That's funny, I was just thinking that keeping track of all these new developments would be a good job for an llm + langchain.  :-)

Alan Downing

unread,
Oct 6, 2023, 2:02:42 AM10/6/23
to Alan Timm, RSSC-List, hbrob...@googlegroups.com
If you want to keep up with the rapid developments in AI, I highly recommend the last-week-in-ai podcast:


Alan

Sent from my iPhone

On Oct 6, 2023, at 1:07 PM, Alan Timm <gest...@gmail.com> wrote:

That's funny, I was just thinking that keeping track of all these new developments would be a good job for an llm + langchain.  :-)

Chris Albertson

unread,
Nov 5, 2023, 5:57:22 PM11/5/23
to Alan Timm, RSSC-List

“LLM” has become too large of an area for one person to follow.   You have to specialize.  Well, you have to specialize if you want to stay at the cutting edge.   

If you want to keep abreast of the wider technology you should be reading review papers.  Some authors are very good and summarizing and comparing recent research.    Or even go with textbooks.  It takes a while for recent technology to find its way into a book.  But this filter-of-time is a good thing.  Well, a good thing unless you are doing research, and then you’d be very specialized

I have two robot-related interests right now (1) MPC (model predictive control) This is the thing everyone should look at after they understand basic PID controllers and the textbook and overview papers and GitHub examples are good for learning. and (2) Foothold selection — how should a robot choose the exact location to place its foot on the ground, given the need to maintain balance, move quickly to its goal, and not slip on uneven ground?  Every footstep is a compromise so there is no good answer.   I think MPC might offer a solution.  This is my current interest but, I know that reading all the papers published on this would be the WORST way to learn.

LLMs are not unique.  Pick any subject in robotics.  Say “Field Oriented Control” for BLDC motors.   The reviews, tutorials and examples are much more usfull then current research.





Reply all
Reply to author
Forward
0 new messages