Revision History
AdamW · 6 revisions
Sizes are character counts of the article source. The signed number is the change from the previous revision.
Recent edit summaries
Detailed summaries recorded by editors. Generic maintenance summaries are omitted here; every recorded revision remains below. Dates describe the edit, not necessarily the event it covers.
- Correction: Corrected a fabricated GPT-2 hyperparameter row and a composite ViT-Large row, reversed the inverted weight-decay scaling rule, fixed the 70B-model optimizer-memory arithmetic, removed unsupported claims that GPT-4, Claude, Gemini and PaLM were trained with AdamW (PaLM used Adafactor), and corrected Muon's memory and source attribution (plus 15 minor fixes)
Version 6 · Nov 1, 2026, 12:31 PM
- Improve internal links (2026-10-13)
Version 5 · Oct 13, 2026, 10:17 AM
- Added 6 contextual internal links
Version 4 · Jul 23, 2026, 03:28 PM
- Render formulas with LaTeX math notation
Version 3 · Jul 13, 2026, 11:05 PM