V2 fixed
From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)
V2 fixed has 33 facts recorded in Dontopedia across 2 references, with 3 live disagreements.
Mostly:lacks observation of(3), has best bpb(2), has final bpb(2)
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.
assertsNoIssuesAsserts No Issues(1)
- Xenonfun
ex:xenonfun
evaluatesPositivelyEvaluates Positively(1)
- Xenonfun
ex:xenonfun
hasSubjectHas Subject(1)
- Comparison V1 Vs V2
ex:comparison-v1-vs-v2
Other facts (32)
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 |
|---|---|---|
| Lacks Observation of | 35k Step Cliff | [1] |
| Lacks Observation of | Catastrophic Forgetting | [1] |
| Lacks Observation of | Gnorm Explosion | [1] |
| Has Best Bpb | 1.722 | [1] |
| Has Best Bpb | 1.722 | [2] |
| Has Final Bpb | 1.841 | [1] |
| Has Final Bpb | 1.841 | [2] |
| Quality Gain Caused by | Fixed Grad Z Math | [1] |
| Quality Gain Caused by | Full Adam Moment Resume | [1] |
| Has Quality Gain Origin | fixed grad_z math | [2] |
| Has Quality Gain Origin | full Adam-moment resume | [2] |
| Achieves Throughput | 44000 | [1] |
| Avoids Buggy Steps | 52K fixed | [1] |
| Best Bpb Improvement Over | V1 Rotadam | [1] |
| Comprises Steps | 10K buggy + 52K fixed | [1] |
| Head to Head Compared With | V1 Rotadam | [1] |
| Is Fixed Version | true | [1] |
| Superior to | V1 Rotadam | [1] |
| Throughput Gain From | O(T·d) vs O(T·d²) structural win (S2-free) | [1] |
| Throughput Multiplier Over | V1 Rotadam | [1] |
| Uses Steps | 62000 | [1] |
| Wins on Axes | both | [1] |
| Rdf:type | Model Version | [2] |
| Has Bpb Difference Relative to V1 | -0.3 | [2] |
| Has Throughput | 44K tok/s | [2] |
| Has Throughput Multiplier | 1.77 | [2] |
| Has Step Count | 62000 | [2] |
| Has Buggy Step Count | 10000 | [2] |
| Has Fixed Step Count | 52000 | [2] |
| Wins on | both axes | [2] |
| Has Throughput Gain Origin | O(T·d) vs O(T·d²) structural win | [2] |
| Is S2 Free | true | [2] |
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 (2)
ctx:discord/blah/watt-activation/part-664ctx:discord/blah/watt-activation/661- full textwatt-activation-661text/plain2 KB
doc:agent/watt-activation-661/b1cbb45f-4fa8-4281-bf4f-564e7c2ebe63Show excerpt
[2026-04-20 00:14] xenonfun: ``` ⏺ Run complete. Task #5 done. Final head-to-head: ┌────────────┬────────────┬─────────────────────────────┐ │ │ V1 rotadam │ V2 fixed │ ├────────────┼────────────┼────…
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.