Revision History
Llama 3.1 · 11 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 Patrick McGuinness quote, a misattributed "three-stage" post-training description (the paper reports six rounds), wrong Llama 3 8B baseline scores, wrong GPT-4o/Claude MMLU-Pro and MBPP figures, an unsupported HellaSwag row, the MATH gap claim, an unsupported Llama 3.3 "one-fifth the inference cost" figure, the quality-classifier training description, the 54-day pre-training claim, the KV-cache comparison, an unsupported licensing-template claim, and updated Llama Stack's 2026 rename to OGX (plus 15 minor fixes)
Version 11 · Nov 2, 2026, 05:22 AM
- Update: added the developer's official video.
Version 10 · Nov 1, 2026, 06:41 AM
- Improve internal links (2026-10-13)
Version 8 · Oct 13, 2026, 10:21 AM
- Added 6 contextual internal links
Version 7 · Jul 23, 2026, 03:29 PM
- Pre-update snapshot before enhancement to 5599 words.
Version 2 · May 17, 2026, 10:37 PM