Revision History
DPO · 12 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 fabricated PPO-vs-DPO benchmark margins, a fabricated quote attributed to the 'It Takes Two' GRPO paper, invented RLHF compute and enterprise-adoption statistics, the wrong institution for a 2025 RLHF-vs-DPO study, the misstated motivation of IPO and the Tulu 3 PPO-vs-DPO finding, and unsupported claims that Llama 3.3, Qwen 3, Gemma 2 and Starling were trained with DPO; also fixed several timeline dates and venues (plus 44 minor fixes)
Version 12 · Nov 2, 2026, 03:27 PM
- Update: added a video.
Version 11 · Nov 1, 2026, 08:25 AM
- Added 6 contextual internal links
Version 10 · Jul 23, 2026, 03:03 PM
- Render formulas with LaTeX math notation
Version 9 · Jul 11, 2026, 10:31 PM