Dontopedia
Explore

Tokens Per Step

From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-07-18.)

Tokens Per Step has 6 facts recorded in Dontopedia across 2 references.

6 facts·6 predicates·2 sources

Mostly:quadruples with seq(1), related to sequence length(1), related to batch size(1)

Maturity scale raw canonical shape-checked rule-derived certified

Quadruples With SeqquadruplesWithSeq

Related to Sequence LengthrelatedToSequenceLength

Related to Batch SizerelatedToBatchSize

Tokens Per Step ValuetokensPerStepValue

  • 4096[2]sourceall time · 0434182b F210 4ea2 Becd 40cbf8e40dd7

Rdfs:labelrdfs:label

  • Tokens per step[2]sourceall time · 0434182b F210 4ea2 Becd 40cbf8e40dd7

Rdf:typerdf:type

Inbound mentions (2)

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.

rdf:typeRdf:type(1)

unitUnit(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.

6 undated
—
quadruplesWithSeqblah/watt-activation/part-355
ex:longer-seq
—
labeldocument/0434182b-f210-4ea2-becd-40cbf8e40dd7
Tokens per step
—
typedocument/0434182b-f210-4ea2-becd-40cbf8e40dd7
ex:tokens-per-step
—
relatedToBatchSizedocument/0434182b-f210-4ea2-becd-40cbf8e40dd7
ex:batch-size-16
—
relatedToSequenceLengthdocument/0434182b-f210-4ea2-becd-40cbf8e40dd7
ex:seq-len-256
—
tokensPerStepValuedocument/0434182b-f210-4ea2-becd-40cbf8e40dd7
4096

References (2)

2 references
  1. [1]Part 3551 fact
    customctx:discord/blah/watt-activation/part-355
  2. [2]xenonfun: ``` At step 2K with batch=16 and seq_len=256: - Tokens per step: 16 × 256 =5 facts
    customctx:claims/document/0434182b-f210-4ea2-becd-40cbf8e40dd7
    • full textxenonfun: ``` At step 2K with batch=16 and seq_len=256: - Tokens per step: 16 × 256 =
      text/plain366 Bdiscord:convmsg/11763
      Show excerpt
      xenonfun: ``` At step 2K with batch=16 and seq_len=256: - Tokens per step: 16 × 256 = 4,096 - Total tokens through 2K steps: 2,000 × 4,096 = 8,192,000 tokens = ~8.2 MB of data The training corpus was 35 MB total, so at step 2K…

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.

↑