Wire
From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)
Wire has 51 facts recorded in Dontopedia across 16 references, with 5 live disagreements.
51 facts·43 predicates·16 sources·5 in dispute
Mostly:rdf:type(4), replaces(2), uses(2)
Maturity scale
raw canonical shape-checked rule-derived certifiedInbound mentions (8)
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.
isBetterAttentionVariantForIs Better Attention Variant for(2)
- Lohe Complex
ex:lohe-complex - Lohe Complex
ex:lohe-complex
contrastsWithContrasts With(1)
- No Symbol Formation Needed
ex:no-symbol-formation-needed
implementsImplements(1)
- Software Defined Radio
ex:software-defined-radio
isIs(1)
- Research Direction 1
ex:research-direction-1
mentionsTopicMentions Topic(1)
- Log Entry 1
ex:log-entry-1
observesWorkingOnObserves Working on(1)
- Xenonfun
ex:xenonfun
syncsAgainstSyncs Against(1)
- First Block
ex:first-block
Other facts (49)
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.
49 facts
| Predicate | Value | Ref |
|---|---|---|
| Rdf:type | Feature | [11] |
| Rdf:type | Concept | [13] |
| Rdf:type | Encoding Method | [15] |
| Rdf:type | Encoding Method | [16] |
| Replaces | Learned Embedding Table | [3] |
| Replaces | Tokenizers | [9] |
| Uses | Zero Learned Encoding Plus Gamma Cumsum | [8] |
| Uses | Qpsk Phase Modulated Byte Representations | [9] |
| Modulation Type | Qpsk | [11] |
| Modulation Type | Quaternion | [11] |
| Superior to Vq | in efficiency | [1] |
| Completely Changed the Game | Model Training | [1] |
| Enables Focus on Structure | Language Model Capacity | [1] |
| Starts With Perfect Symbols | true | [1] |
| Overcomes Disadvantage of Byte Prediction | null | [2] |
| Maintains Ontological Bit Structure | null | [2] |
| Is Slightly Better on Per Byte Ce | null | [2] |
| Is Competitive on Per Byte Ce | null | [2] |
| Has Geometric Structure | true | [4] |
| Presupposes Geometric Structure | Structured | [5] |
| Contextual in Physics ML | Geometric Structured Input | [5] |
| Is | Qpsk Plus Hamming Plus Haar | [6] |
| Maps | Byte to Phasor on S 14h 1 | [6] |
| Has Compute Complexity | O(1) lookup | [7] |
| Has Memory Complexity | O(T × G × H) = T × 16 floats | [7] |
| Improves With Scale | Ppl Reduction | [8] |
| At Params | 14K | [8] |
| Achieves Ppl | 9.17 | [8] |
| Scales to Ppl | 3.92 | [8] |
| Scales at Params | 2.3M | [8] |
| Uses Lohe Sync As | Iterative Decoding | [9] |
| Is Fascinating | Direction | [9] |
| Is Most Novel | true | [9] |
| Uses Fec As | Parity Oscillators | [9] |
| Starts With | 256 Symbol Constellation | [10] |
| Operation | Byte to Phase Modulation | [11] |
| Includes | Fec | [11] |
| Bpb Score | 2.39 | [12] |
| Parameter Count | 3930000 | [12] |
| Training Duration | 39 seconds | [12] |
| Superior to | Vq Encoding | [12] |
| Bpb Advantage Factor | 2 | [12] |
| Starts Optimal | true | [12] |
| Skips Symbol Formation Phase | true | [12] |
| Training Steps | 1000 | [12] |
| Has Property | standard | [13] |
| Game Changing Status | completely changed the game | [14] |
| Initial State | perfect symbols | [14] |
| Effect on Model | lets the model spend all its capacity on learning language structure | [14] |
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.
—
superiorToVqblah/watt-activation/part-326
in efficiency
—
completelyChangedTheGameblah/watt-activation/part-326
ex:model-training
—
enablesFocusOnStructureblah/watt-activation/part-326
ex:language-model-capacity
—
startsWithPerfectSymbolsblah/watt-activation/part-326
true
—
overcomesDisadvantageOfBytePredictionblah/watt-activation/part-327
null
—
maintainsOntologicalBitStructureblah/watt-activation/part-327
null
—
isSlightlyBetterOnPerByteCeblah/watt-activation/part-327
null
—
isCompetitiveOnPerByteCeblah/watt-activation/part-327
null
—
replacesblah/watt-activation/part-333
ex:learned-embedding-table
—
hasGeometricStructureblah/watt-activation/part-350
true
—
presupposesGeometricStructureblah/watt-activation/part-351
ex:structured
—
contextualInPhysicsMlblah/watt-activation/part-351
ex:geometric-structured-input
—
isblah/watt-activation/part-378
ex:qpsk-plus-hamming-plus-haar
—
mapsblah/watt-activation/part-378
ex:byte-to-phasor-on-s-14h-1
—
hasComputeComplexityblah/watt-activation/part-401
O(1) lookup
—
hasMemoryComplexityblah/watt-activation/part-401
O(T × G × H) = T × 16 floats
—
improvesWithScaleblah/watt-activation/part-491
ex:ppl-reduction
—
usesblah/watt-activation/part-491
ex:zero-learned-encoding-plus-gamma-cumsum
—
atParamsblah/watt-activation/part-491
14K
—
achievesPplblah/watt-activation/part-491
9.17
—
scalesToPplblah/watt-activation/part-491
3.92
—
scalesAtParamsblah/watt-activation/part-491
2.3M
—
usesLoheSyncAsblah/watt-activation/part-320
ex:iterative-decoding
—
isFascinatingblah/watt-activation/part-320
ex:direction
—
isMostNovelblah/watt-activation/part-320
true
—
usesblah/watt-activation/part-320
ex:qpsk-phase-modulated-byte-representations
—
replacesblah/watt-activation/part-320
ex:tokenizers
—
usesFecAsblah/watt-activation/part-320
ex:parity-oscillators
—
startsWithblah/watt-activation/part-325
ex:256-symbol-constellation
—
typeblah/watt-activation/316
ex:Feature
—
modulationTypeblah/watt-activation/316
ex:QPSK
—
modulationTypeblah/watt-activation/316
ex:quaternion
—
operationblah/watt-activation/316
ex:byte-to-phase-modulation
—
includesblah/watt-activation/316
ex:FEC
—
labelblah/watt-activation/321
Wire
—
bpbScoreblah/watt-activation/321
2.39
—
parameterCountblah/watt-activation/321
3930000
—
trainingDurationblah/watt-activation/321
39 seconds
—
superiorToblah/watt-activation/321
ex:vq-encoding
—
bpbAdvantageFactorblah/watt-activation/321
2
—
startsOptimalblah/watt-activation/321
true
—
skipsSymbolFormationPhaseblah/watt-activation/321
true
—
trainingStepsblah/watt-activation/321
1000
—
typeblah/watt-activation/318
ex:Concept
—
labelblah/watt-activation/318
wire encoding
—
hasPropertyblah/watt-activation/318
standard
—
gameChangingStatusblah/watt-activation/324
completely changed the game
—
initialStateblah/watt-activation/324
perfect symbols
—
effectOnModelblah/watt-activation/324
lets the model spend all its capacity on learning language structure
—
typeblah/watt-activation/323
ex:EncodingMethod
—
typeblah/watt-activation/335
ex:EncodingMethod
References (16)
16 references
ctx:discord/blah/watt-activation/part-326ctx:discord/blah/watt-activation/part-327ctx:discord/blah/watt-activation/part-333ctx:discord/blah/watt-activation/part-350ctx:discord/blah/watt-activation/part-351ctx:discord/blah/watt-activation/part-378ctx:discord/blah/watt-activation/part-401ctx:discord/blah/watt-activation/part-491ctx:discord/blah/watt-activation/part-320ctx:discord/blah/watt-activation/part-325ctx:discord/blah/watt-activation/316- full textwatt-activation-316text/plain3 KB
doc:agent/watt-activation-316/c4f50044-9de8-4f34-94c9-40d82203b46bShow excerpt
[2026-03-15 01:57] xenonfun: ``` rot_strength = 0.1 (new default) ┌────────────────┬──────┬─────┬────────┬────────┐ │ Attention │ Loss │ PPL │ Tok/s │ Params │ ├────────────────┼──────┼─────┼────────┼────────┤ │ spectral …
ctx:discord/blah/watt-activation/321- full textwatt-activation-321text/plain2 KB
doc:agent/watt-activation-321/7ca21bd7-e2e0-473c-bf6a-7562309553b0Show excerpt
[2026-03-15 03:25] xenonfun: ``` ⏺ Summary Wire vs VQ: ┌──────────┬───────────────────────┬──────────────────────┐ │ │ Wire (best, 1K steps) │ VQ (best, 20K steps) │ ├──────────┼───────────────────────┼───────────────────…
ctx:discord/blah/watt-activation/318- full textwatt-activation-318text/plain3 KB
doc:agent/watt-activation-318/f52d95a8-f461-40d1-9360-f08558b18eb1Show excerpt
[2026-03-15 02:47] xenonfun: ⏺ I see you're working on wire encoding / phase modulation — that's a fascinating direction. Let me check what you've got: [2026-03-15 02:47] lisamegawatts: Wire QPSK + Standard: PPL 4.94, Byte Accuracy 51.5% T…
ctx:discord/blah/watt-activation/324- full textwatt-activation-324text/plain2 KB
doc:agent/watt-activation-324/5428b43d-e96d-4178-8924-6d7c75574aa4Show excerpt
[2026-03-15 03:48] xenonfun: ## Design (files: Screenshot_2026-03-14_at_11.48.03_PM.png) [2026-03-15 03:49] xenonfun: (files: Screenshot_2026-03-14_at_11.49.32_PM.png) [2026-03-15 04:03] xenonfun: You went from 2.39 → 1.572 today. You cros…
ctx:discord/blah/watt-activation/323- full textwatt-activation-323text/plain3 KB
doc:agent/watt-activation-323/60722a8e-0dbe-4aee-be7b-862e43a25b6fShow excerpt
[2026-03-15 03:41] xenonfun: ``` Elevator Pitch We built a language model that doesn't use a tokenizer. Instead of BPE or word pieces, we encode raw bytes directly as phase-modulated oscillators using digital communications theory…
ctx:discord/blah/watt-activation/335- full textwatt-activation-335text/plain2 KB
doc:agent/watt-activation-335/dd79a73c-09f4-4ca4-992c-61e51914a192Show excerpt
[2026-03-15 14:31] xenonfun: ⏺ 8.40M params — essentially the same as the lohe_spherical run (8.45M). Same model size, same data shape, but lohe_complex + SEQ=256 reached ppl=6.1 in 855 steps vs 40,000 steps. That's a ~47× faster converg…
See also
- Model Training
- Language Model Capacity
- Learned Embedding Table
- Structured
- Geometric Structured Input
- Qpsk Plus Hamming Plus Haar
- Byte to Phasor on S 14h 1
- Ppl Reduction
- Zero Learned Encoding Plus Gamma Cumsum
- Iterative Decoding
- Direction
- Qpsk Phase Modulated Byte Representations
- Tokenizers
- Parity Oscillators
- 256 Symbol Constellation
- Feature
- Qpsk
- Quaternion
- Byte to Phase Modulation
- Fec
- Vq Encoding
- Concept
- Encoding Method
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.