Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up
RakshitAralimatti 
posted an update 7 days ago
Post
6709
When you ask ChatGPT, Claude, or Gemini a really tough question,
you might notice that little "thinking..." moment before it answers.

But what does it actually mean when an LLM is “thinking”?

Imagine a chess player pausing before their next move not because they don’t know how to play, but because they’re running through possibilities, weighing options, and choosing the best one.
LLMs do something similar… except they’re not really thinking like us.

Here’s the surprising part :-
You might think these reasoning skills come from futuristic architectures or alien neural networks.
In reality, most reasoning LLMs still use the same transformer decoder-only architecture as other models
The real magic?
It’s in how they’re trained and what data they learn from.

Can AI actually think, or is it just insanely good at faking it?
I broke it down in a simple, 4-minute Medium read.
Bet you’ll walk away with at least one “aha!” moment. 🚀

Read here - https://lnkd.in/edZ8Ceyg

Outstanding!!

·

Thanks @Edalexan .Glad that you found it useful.

Not a bad start, but it needs compression and doesn't include examples or quantitative justifications for the areas that benefit from reasoning. Doesn't even mention backtracking, which is something in reasoning models I'd like to learn more about. What are examples of "novel" reasoning paths?

What was the research paper which introduced reasoning? You don't list any sources or external links.

·

Thanks @treehugg3 for the thoughtful feedback! 🙏 My main motive here was to break it down in simple words for people who are new to AI or just starting to learn about reasoning models. I completely understand your concern I’ll make sure to include more detailed explanations, examples, sources, and technical depth (like backtracking and novel reasoning paths) in upcoming blogs. Really appreciate your input, it helps me improve!

This is one of my primary study directions - not for the intent to produce "thinking" but with the intent to reproduce the very essence of "one-ness" that manifests as an eventuality when large corpus are trained on LLMS and then curated using specific paradigms in careful ways.
Imposing this behavior onto smaller models will hopefully allow the very curative potential of self-regulated intelligence that will require less curation and more conversation.
I'm exploring many variants of geometry to impose this exact behavior - with the end result goal of giving considerably smaller models the ability to understand how to pilot themselves, without needing 200 billion params.

A number of models i wish i could get it to shut up on the thinking, in fact 'thinking' modes tends to turn me off on what would otherwise be good models.

Thinking generally is a summarizing of all the details, sometimes giving better insight on what has more emphasis, but usually the 'thinking' i see is useless, as half the output beyond being a summary it talks about staying within guidelines and not offending anyone. 99% of the time 'thinking' is just a waste of time.

Then again i'm not using the cloud versions so maybe those actually are much better at it.


Actually one thing 'thinking' might do for very very large models, is to gather appropriate data all in one block right before it answers. Thus it might use 2 modes, a very large context mode, and then a shorter faster one with the more recent context. That in theory may work...

It is there for a long time. Listen for the clicks. That is when you’re in a good chat.

Excellent read! Thanks for the insights here.

And thanks for the follow! (Harmonic Frontier Audio - Scottish Smallpipes Audio Dataset Preview) Great to connect!