Viewing profile — mesuvash

IIUC, The paper's notation S^(d-1) means the unit sphere in R^d (e.g., the familiar unit circle is S^1 living in R^2). So, i think, x in the algorithm is already a unit vector. Ref…

Fair point. I've updated the animation to address this. The grid now uses the correct non-uniform centroids (optimal for the arcsine distribution in 2D), so you'll see grid lines c…

Yes, this is important in high dimension. But sadly, very hard to visualize. In 2d it looks like unnecessary.

That's actually correct and intentional. TurboQuant applies the same rotation matrix to every vector. The key insight is that any unit vector, when multiplied by a random orthogona…

Yes. Great catch. I simplified the grid just for visualization purpose. I've updated the visualization. The grid is actually not uniformly spaced. Each coordinate is quantized inde…

Author here. Sorry still working on refining the post. Will share once the post is ready.

TurboQuant explained with an easy to understand (no-math) animation https://mesuvash.github.io/blog/2026/turboquant-interactive/

I am glad you liked it :) You might like this https://mesuvash.github.io/blog/2026/rl_for_llm/ as well :)

Thanks for the pointers. From my personal experience, Auto-encoders are amazing for dense input (images, audio etc), more specifically, when the input feature space is not large. H…

>>I think the hashes could take some work. Any suggestions or thing that are not clear? Thanks for your feedback. I shall update the post accordingly.

Nothing but #respect. It's hard to see people who give up fortune for what they consider right thing(atleast from his perspective).

Seaborn is a library of high-level functions that facilitate making informative and attractive plots of statistical data using Matplotlib.