Eval Table
From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)
Eval Table has 6 facts recorded in Dontopedia across 1 reference, with 1 live disagreement.
6 facts·3 predicates·1 sources·1 in dispute
Maturity scale
raw canonical shape-checked rule-derived certifiedOther facts (6)
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.
6 facts
| Predicate | Value | Ref |
|---|---|---|
| Has Column | Model Column | [1] |
| Has Column | Loss Column | [1] |
| Has Column | Ppl Column | [1] |
| Has Column | Delta Ppl Column | [1] |
| Rdf:type | Table | [1] |
| Has Dataset Name | philosophy_finetune/val.txt | [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.
—
typeblah/watt-activation/25
ex:Table
—
hasColumnblah/watt-activation/25
ex:model-column
—
hasColumnblah/watt-activation/25
ex:loss-column
—
hasColumnblah/watt-activation/25
ex:ppl-column
—
hasColumnblah/watt-activation/25
ex:delta-ppl-column
—
hasDatasetNameblah/watt-activation/25
philosophy_finetune/val.txt
References (1)
1 references
ctx:discord/blah/watt-activation/25- full textwatt-activation-25text/plain2 KB
doc:agent/watt-activation-25/cd4b2fb2-f4ea-4f95-b51a-eb3eeadbe176Show excerpt
[2026-03-06 15:40] xenonfun: New best 1.0445 — each warm restart cycle is pushing it lower (1.6636 → 1.1405 → 1.1394 → 1.0445). The pattern is consistent: best always hits around iter ~1450 during LR warmup on some easy batch, then the run …
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.