Show HN: Microgpt is a GPT you can visualize in the browser

(microgpt.boratto.ca)

268 points | by b44 1 day ago

16 comments

  • kengoa 16 hours ago
    Amazing work! Reminded me of LLM Visualization (https://bbycroft.net/llm) except this is a lot easier to wrap my head around and that I can actually run the training loops, which makes sense given the simplicity of the original microgpt.

    To give a sense of what the loss value means, maybe you can add a small explainer section as a question and add this explanation from Karpathy’s blog:

    > Over 1,000 steps the loss decreases from around 3.3 (random guessing among 27 tokens: −log(1/27)≈3.3) down to around 2.37.

    to reiterate that the model is being trained to predict the next token out of 27 possible tokens and is now doing better than the baseline of random guess.

  • interloxia 16 hours ago
    The linked inspiration has a good blog post of microgpt implemented in python.

    https://karpathy.github.io/2026/02/12/microgpt/

    It was submitted to hn a few days ago but only received a few comments. https://news.ycombinator.com/item?id=47000263

  • krackers 1 day ago
    There used to be this page that showed the activations/residual stream from gpt-2 visualized as a black-white image. I remember it being neat how you could slowly see order forming from seemingly random activations as it progressed through the layers.

    Can't find it now though (maybe the link rotted?), anyone happen to know what that was?

  • RugnirViking 16 hours ago
    I was a little confused by "see, its much better" when the output is stuff like isovrak and kucey. What is it supposed to be generating?
    • b44 15 hours ago
      the untrained model is literally just generating random characters, whereas your examples are at least pronouncable. you can add more layers to get progressively better results.
    • lucrbvi 16 hours ago
      It's just hallucinating training data, the model is very small so it cannot be useful at all
  • stevage 21 hours ago
    I'd love to understand how LLMs work, but this site assumed a bit too much knowledge for me to get much from it. Looks cool though.
  • armcat 4 hours ago
    Really nicely presented, well done!
  • prakashdep 7 hours ago
    It reminds me the anything+GPT era of 2022-2024
  • ramon156 15 hours ago
    My Android phone was not a fan of this site, but on my desktop it works great! Cool stuff
  • umairnadeem123 20 hours ago
    nice. visualizing the prompt->tool->output graph is underrated, it makes failure modes (and cost) obvious. do you track token/call cost per node + cache hits, or is it purely structural right now? also curious if you let users diff two runs (same prompt, different model/tool) and see which node diverged first.
  • keepamovin 15 hours ago
    I can't help but think there has to be a cheaper way to LLM.
  • kfsone 1 day ago
    Minor nit: In familiarity, you gloss over the fact that it's character rather than token based which might be worth a shout out:

    "Microgpt's larger cousins using building blocks called tokens representing one or more letters. That's hard to reason about, but essential for building sentences and conversations.

    "So we'll just deal with spelling names using the English alphabet. That gives us 26 tokens, one for each letter."

    • mips_avatar 1 day ago
      Using ascii characters is a simple form of tokenization with less compression
    • b44 1 day ago
      hm. the way i see things, characters are the natural/obvious building blocks and tokenization is just an improvement on that. i do mention chatgpt et al. use tokens in the last q&a dropdown, though
  • msla 1 day ago
    About how many training steps are required to get good output?
    • alansaber 15 hours ago
      Depends on the model size, batch size, input sequence length, ... etc. With a small model like this you'll never get a 'good' output but you can maximise its potential.
    • WatchDog 1 day ago
      I trained 12,000 steps at 4 layers, and the output is kind of name-like, but it didn't reproduce any actual name from it's training data after 20 or so generations.
    • b44 1 day ago
      not many. diminishing returns start before 1000 and past that you should just add a second/third layer
  • GaggiX 20 hours ago
    Wtok and Wpos should be 26-dim along one of the axis but it shows a 16x16 matrix be default, fc1 instead 16x64 with the default settings (not 16x16).
    • b44 20 hours ago
      good catch - i intentionally cap node visualizations at 16 so it doesn't get super long, but the sidebar shouldn't have that
  • youio 15 hours ago
    really well done
  • darepublic 1 day ago
    thank you for this
  • nivcmo 15 hours ago
    [dead]