Revision History
Training run · 7 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.
- Update: capitalize PaLM link labels
Version 7 · Nov 5, 2026, 12:16 PM
- Correction: Corrected PaLM's training duration (about 64 days, not 50), OPT-175B's restart count (at least 35 manual restarts and over 100 hosts cycled, not over 100 restarts), Llama 3's batch-size ramp (starting at 4M tokens), the 2.78-hour figure (all interruptions, not hardware failures), removed an unsupported claim about sub-$1B runs, and noted that the 2023 US AI executive order was revoked in January 2025 (plus 19 minor fixes)
Version 6 · Nov 2, 2026, 05:35 PM
- Improve internal links (2026-10-13)
Version 5 · Oct 13, 2026, 10:06 AM
- Added 6 contextual internal links
Version 4 · Jul 23, 2026, 03:45 PM