cd /news/artificial-intelligence/llm-city-3d-render-of-all-kimi-k3-s-… · home topics artificial-intelligence article
[ARTICLE · art-100022] src=magik.net ↗ pub= topic=artificial-intelligence verified=true sentiment=· neutral

LLM City – 3D render of all Kimi K3's weights as 2.5mm tiles

A new interactive 3D visualization, LLM City, renders all weights of Kimi K3 as 2.5mm tiles, with X and Y axes representing matrix dimensions and Z representing execution depth, preserving exact matrix areas while allowing 896 experts to exist in parallel. The tool includes runtime cache overlays, router choice visualizations, and flight mode with collision detection, offering a detailed view of the model's architecture.

read2 min views4 publishedAug 17, 2026

3D was chosen deliberately. X and Y are the two matrix axes, so a tensor with shape [out,in] is drawn as an in × out rectangle. Z is execution depth. This keeps every matrix area exact while giving the 896 experts room to exist in parallel.

The physical atlas rotates some rectangles by 90° only to pack them tightly; area remains exact. At the default world scale, every scalar is a 2.5 mm × 2.5 mm tile. The unit selector can deliberately rescale that physical interpretation.

Runtime caches are overlays in both views. One cache scalar remains one square scene unit. MLA key/value caches are split into 32,768-token pages and folded through depth; KDA preserves all 96 recurrent head states plus q/k/v short-convolution history in exact-area aggregate sheets.

Router choices are input-dependent. Bright top-16 lines use a deterministic illustrative set of experts for the active layer; the faint fan-out shows all 896 candidates.

Controls

Left drag: orbit. Right drag or Shift+drag: pan. Wheel/pinch: percentage dolly toward the exact surface under the pointer. A hovered/selected surface becomes a hard 12 cm camera barrier.

Flight mode: drag to look, W/A/S/D to move, E/Q to change altitude, and Shift to sprint. Touch devices get independent move and look joysticks plus altitude and boost controls. Swept collision against learned matrices and visible cache sheets provides wall sliding at a 12 cm standoff. The ground does not clamp flight altitude.

Click a matrix to select it. Double-click focuses it.

F fit all, L active layer, S selected, G toggle flight/orbit, Space outside walking mode.

Every matrix and cache sheet carries procedural screen-space text. At close range the renderer shows the literal scalar grid plus mock digits. Once scalars go sub-pixel, it transitions to a stochastic footprint filter so zoomed-out values read as tinted white noise instead of moire stripes; the full glyphs continuously resolve again as you approach. These are visual placeholders, not checkpoint values.

Selecting a tensor highlights its downstream matrices and related operation nodes; the inspector lists clickable connections.

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @kimi k3 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/llm-city-3d-render-o…] indexed:0 read:2min 2026-08-17 ·