Revision History
Ring Attention · 9 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 the A100 minimum block size (about 1,000 tokens, not 256-1,024), the 30B TPUv4-1024 result (2M tokens, not over 100 million), false claims that DeepSpeed and PyTorch Lightning ship Ring Attention hybrids, misstatements of Magic's LTM-2-mini blog (not the first commercial 100M model, no mention of Ring Attention), and the GPU MODE lecture date (April 2024) (plus 18 minor fixes)
Version 9 · Nov 4, 2026, 10:02 PM
- Improve internal links (2026-10-13)
Version 8 · Oct 13, 2026, 10:04 AM
- Added 6 contextual internal links
Version 7 · Jul 23, 2026, 03:31 PM
- Render formulas with LaTeX math notation
Version 6 · Jul 13, 2026, 11:05 PM
- Initial AI-authored creation (4664 words).
Version 2 · May 18, 2026, 01:04 PM