Revision History
The Pile (dataset) · 10 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 Pile's token count (about 300-340 billion, not 1.35 trillion), reversed a false claim that the released dataset was decontaminated against benchmarks, fixed the claim that the OpenAI lawsuit alleged Books3 use, corrected Shawn Presser's stated rationale and the datasheet's subpopulation statement, and re-sourced the Books3 takedown details (plus 20 minor fixes)
Version 10 · Nov 2, 2026, 03:34 PM
- Improve internal links (2026-10-13)
Version 9 · Oct 13, 2026, 10:09 AM
- Added 6 contextual internal links
Version 8 · Jul 23, 2026, 03:06 PM