spherical VQ
From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)
spherical VQ is vector quantization on spheres.
Mostly:has performance concern(6), has good pattern(4), includes component(3)
Maturity scale
raw canonical shape-checked rule-derived certifiedInbound mentions (14)
Other subjects in dontopedia point AT this entity as a value. These are inverse relationships — e.g. "X motherOf this subject" — and answer questions the forward facts can't. Grouped by predicate.
containsComponentContains Component(2)
- User Architecture
ex:user-architecture - User Pipeline
ex:user-pipeline
advocatesReplacingAdvocates Replacing(1)
- Xenonfun
ex:xenonfun
existentiallyCommittedToExistentially Committed to(1)
- 32 Codes
ex:32-codes
hasItemHas Item(1)
- Open Items List
ex:open-items-list
isIs(1)
- Pipeline Stage 3
ex:pipeline-stage-3
isTargetFrameworkIs Target Framework(1)
- Mlx
ex:mlx
lacksDocumentationForLacks Documentation for(1)
- Portal
ex:portal
outputsToOutputs to(1)
- Lohespherical Encoder
ex:lohespherical-encoder
precedesStagePrecedes Stage(1)
- Lohe Spherical Encoder
ex:lohe-spherical-encoder
providesMlXPerformanceReviewForProvides ML X Performance Review for(1)
- Xenonfun
ex:xenonfun
ranksAsSecondOptionRanks As Second Option(1)
- Xenonfun
ex:xenonfun
stage3Stage3(1)
- Processing Pipeline
ex:processing-pipeline
transformsIntoTransforms Into(1)
- Harmonic Encoder
ex:harmonic-encoder
Other facts (40)
The long tail: predicates that appear too rarely to warrant their own section. Filter or scroll to find a specific one. Each row links to its source.
Timeline
Timeline axis is valid_time — when each source says the fact was true in the world, not when Dontopedia learned about it. Retracted rows are kept for provenance; coloured stripes indicate the context kind.
References (15)
ctx:discord/blah/watt-activation/part-279ctx:discord/blah/watt-activation/part-297ctx:discord/blah/watt-activation/part-300ctx:discord/blah/watt-activation/part-309ctx:discord/blah/watt-activation/part-371ctx:discord/blah/watt-activation/part-370ctx:discord/blah/watt-activation/part-430ctx:discord/blah/watt-activation/part-477ctx:discord/blah/watt-activation/part-302ctx:discord/blah/watt-activation/284- full textwatt-activation-284text/plain2 KB
doc:agent/watt-activation-284/c0d87175-584d-4268-9b87-d373d011109eShow excerpt
[2026-03-13 23:50] xenonfun: Claude: ``` ⏺ Clean takeaway: What works: - Spherical VQ architecture on S^{d-1} - Two-phase training (form codes, then use codes) - Assignment density scaling law (~4 pos/code minimum) - Direction+g…
ctx:discord/blah/watt-activation/298- full textwatt-activation-298text/plain2 KB
doc:agent/watt-activation-298/f5cde311-fd9a-43e7-a746-9177b5a91feeShow excerpt
[2026-03-14 05:53] xenonfun: ``` What Changed The AR decoder produces recognizable English word fragments: "the", "and", "for", "with", "this", "one", "protec(tion)", "earch", "context", "project", "state", "imported". These are real m…
ctx:discord/blah/watt-activation/295- full textwatt-activation-295text/plain3 KB
doc:agent/watt-activation-295/3934680b-d58b-4c73-8470-2c337c1a045eShow excerpt
[2026-03-14 04:39] xenonfun: ```❯ ⏺ Now I have the full picture. Here's my MLX performance review: Spherical VQ — MLX Performance Review Good patterns: 1. _l2_normalize uses + eps inside sqrt (line 38) — matches lohe_normalize sema…
ctx:discord/blah/watt-activation/300- full textwatt-activation-300text/plain3 KB
doc:agent/watt-activation-300/3b6edccf-3524-4608-838f-25890efaea15Show excerpt
[2026-03-14 06:34] xenonfun: ``` 3. Manual attention (lines 110-128) — Hand-rolled softmax attention instead of using mx.fast.scaled_dot_product_attention. MLX's fused attention kernel is significantly faster for small sequence lengths. …
ctx:discord/blah/watt-activation/307- full textwatt-activation-307text/plain2 KB
doc:agent/watt-activation-307/bcc895dd-9148-4cd1-b67f-d61adcb12b77Show excerpt
[2026-03-14 22:38] xenonfun: ```7. Why this matters for your architecture This is the part that is actually useful for you. Your pipeline is roughly: stream → harmonic lift → Lohe dynamics → readout stream→harmonic lift→Lohe dynamics→rea…
ctx:discord/blah/watt-activation/428- full textwatt-activation-428text/plain3 KB
doc:agent/watt-activation-428/2888b7ac-ac11-4017-8748-179fa76f2184Show excerpt
[2026-03-20 02:56] xenonfun: ⏺ Significant gaps — 19 modules and 30 classes missing. The portal covers the high-level architecture but is missing: Major subsystems not documented: 1. Lohe Diffusion — image generation via Lohe manifold …
See also
- Tests for Spherical Vq
- Quantize Euclidean Z Sq Redundant
- Compute Diagnostics Heavy
- Item in Revive Dead Codes
- L2 Normalize Duplicates Lohe Normalize
- Log Norm Gain Mode No Op
- Update Ema Allocates
- Primitives Py
- L2 Normalize Uses Plus Eps Inside Sqrt
- Sampled Separation Penalty
- Ste Via Stop Gradient
- Vectorized Dead Code Revival
- Structural Codes
- Lohe Dynamics Function
- Explicit Codebook
- Anchor Kan
- Mechanism
- Binding Issue
- Entity Addresses
- Sphericalcodebook
- Vector Quantization on Spheres
- Sphericalvqbottleneck
- Sphericalvqhead
- Code Lm
- Architecture
- S D 1 Surface
- Software Component
- Choose Nearest Code
- Lohe Dynamics
- Spherical Codebook
- Spherical Vq Bottleneck
- Spherical Vq Head
Keep researching
Missing something or suspicious of what's here? Kick off a research session — a Claude agent will investigate, cite its sources, and file new facts into a dedicated context you can review before accepting into the shared view.