4 ms·
Show HN: Microgpt is a GPT you can visualize in the browser
very much inspired by karpathy's microgpt of the same name. it's (by default) a 4000 param GPT/LLM/NN that learns to generate names. this is sorta an educational tool in that you can visualize the activations as they pass through the network, and click on things to get an explanation of them.
- kfsone 8mo agoMinor 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."
- b44 8mo agohm. 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
- mips_avatar 8mo agoUsing ascii characters is a simple form of tokenization with less compression
- msla 8mo agoAbout how many training steps are required to get good output?
- b44 8mo agonot many. diminishing returns start before 1000 and past that you should just add a second/third layer
- WatchDog 8mo agoI 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.
- alansaber 8mo agoDepends 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.
- darepublic 8mo agothank you for this
- krackers 8mo agoThere 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?
- umairnadeem123 8mo ago[dead]
- stevage 8mo agoI'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.
- BloondAndDoom 8mo agoThis is the best content I’ve found to to learn how LLMs really work : https://youtu.be/7xTGNNLPyMI?si=Gk0u4suz8pv39tP4 https://youtu.be/7xTGNNLPyMI?si=Gk0u4suz8pv39tP4
- alansaber 8mo agoI think this blog post in particular might be helpful here https://sebastianraschka.com/blog/2023/self-attention-from-scratch.html https://sebastianraschka.com/blog/2023/self-attention-from-s...
- GaggiX 8mo agoWtok 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 8mo agogood catch - i intentionally cap node visualizations at 16 so it doesn't get super long, but the sidebar shouldn't have that
- kengoa 8mo agoAmazing work! Reminded me of LLM Visualization (https://bbycroft.net/llm 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.
- RugnirViking 8mo agoI 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?
- interloxia 8mo agoThe linked inspiration has a good blog post of microgpt implemented in python. https://karpathy.github.io/2026/02/12/microgpt/ 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 https://news.ycombinator.com/item?id=47000263
- ramon156 8mo agoMy Android phone was not a fan of this site, but on my desktop it works great! Cool stuff
- nivcmo 8mo ago[dead]
- youio 8mo agoreally well done
- keepamovin 8mo agoI can't help but think there has to be a cheaper way to LLM.
- prakashdep 8mo agoIt reminds me the anything+GPT era of 2022-2024
- armcat 8mo agoReally nicely presented, well done!