Test 4: HarmonicGPT forward (lohe_spherical+lohe_v2)
From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)
Test 4: HarmonicGPT forward (lohe_spherical+lohe_v2) has 26 facts recorded in Dontopedia across 4 references.
Mostly:rdf:type(2), result success true(1), contrasts with(1)
Maturity scale
raw canonical shape-checked rule-derived certifiedInbound mentions (3)
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.
evaluatedAsSuccessfulEvaluated As Successful(1)
- Location Map
ex:location-map
precedesPrecedes(1)
- Test 3
ex:test-3
testedAsInputTested As Input(1)
- 47 6062 122 3321 Coords
ex:47-6062--122-3321-coords
Other facts (25)
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 |
|---|---|---|
| Rdf:type | Software Test | [3] |
| Rdf:type | Software Test | [4] |
| Result Success True | True | [1] |
| Contrasts With | Test 1 | [1] |
| Provided Args Zoom | 12 | [1] |
| Provided Args Text | 47.6062, -122.3321 | [1] |
| Has Parameter Count | 1,047,920 | [2] |
| Compares to Random Baseline | 6.9315 | [2] |
| Establishes Baseline Loss | Random Baseline | [2] |
| Has Initial Loss | 7.3722 | [2] |
| Has Logits Shape | (4, 128, 1024) | [2] |
| Passes Verification | Checkmark | [2] |
| Tests Forward Pass | Harmonic Gpt | [2] |
| Uses Combination | lohe_spherical+lohe_v2 | [2] |
| Has Status | Success | [3] |
| Has Result | "0000000000000000000000000000000000000000000000000000000000000000:0" | [3] |
| Test Name | HarmonicGPT forward (lohe_spherical+lohe_v2) | [4] |
| Part of Log | Log Entry 1 | [4] |
| Tested Model | Harmonic Gpt | [4] |
| Model Configuration | lohe_spherical+lohe_v2 | [4] |
| Parameter Count | 1047920 | [4] |
| Verified Logits Shape | (4, 128, 1024) | [4] |
| Measured Initial Loss | 7.3722 | [4] |
| Expected Random Baseline Loss | 6.9315 | [4] |
| Test Result | passed | [4] |
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 (4)
ctx:discord/blah/omega/part-708ctx:discord/blah/watt-activation/part-201ctx:discord/blah/general/38- full textgeneral-38text/plain3 KB
doc:agent/general-38/cff650d1-42c4-4389-9e79-0054563c13d0Show excerpt
[2025-04-24 18:14] attackinghobo: also this could be modified to make refactoring really easy [2025-04-24 18:14] attackinghobo: it could make 5 new files with like 10 small lines of xml to copy-paste from one file to another [2025-04-24 18:…
ctx:discord/blah/watt-activation/200- full textwatt-activation-200text/plain2 KB
doc:agent/watt-activation-200/6685d5c1-c0e6-4972-8f0a-8720129da791Show excerpt
[2026-03-10 06:35] xenonfun: LoheSphericalAttention smoke test d_model=256, n_layers=4, G=4, H=4 seq_len=128, batch=4, vocab=1024 ============================================================ Test 1: Forward shapes and r_global output shap…
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.