Scaling Observation
From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)
Scaling Observation has 12 facts recorded in Dontopedia across 1 reference, with 3 live disagreements.
Mostly:explains benefit of more channels(3), states effect of doubling layers(2), explains scope of d model(2)
Maturity scale
raw canonical shape-checked rule-derived certifiedOther facts (12)
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.
| Predicate | Value | Ref |
|---|---|---|
| Explains Benefit of More Channels | more independent signals in the field | [1] |
| Explains Benefit of More Channels | better multi-target retrieval | [1] |
| Explains Benefit of More Channels | richer context transport | [1] |
| States Effect of Doubling Layers | doubles params | [1] |
| States Effect of Doubling Layers | doubles the total stacked propagation distance | [1] |
| Explains Scope of D Model | embedding | [1] |
| Explains Scope of D Model | proj_in/out | [1] |
| Indicates Best First Lever | Channels | [1] |
| Indicates Second Lever | Layers | [1] |
| States Constraint of Layer Increase | without increasing per-block cost | [1] |
| Evaluates Efficiency | D Model | [1] |
| Explains Exclusion of D Model | field physics | [1] |
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 (1)
ctx:discord/blah/watt-activation/353- full textwatt-activation-353text/plain3 KB
doc:agent/watt-activation-353/cc7a24c1-66ae-472e-a74c-30bb70fe2a69Show excerpt
[2026-03-17 09:19] xenonfun: ``` ============================================================ K4_cur10 K=4 curriculum=10% ============================================================ step 1000/5000 BPB=3.173 719,581 tok/s step 2…
See also
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.