Revision History
BIG-Bench Hard · 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.
- Update: regenerate results chart from the fact-checked results table
Version 11 · Nov 4, 2026, 10:34 PM
- Correction: Corrected the per-task results table, whose columns were shifted (human and random baselines were shown as InstructGPT scores and PaLM scores were missing), the Word Sorting claim built on it, a GPT-4 BBH score wrongly attributed to the GPT-4 technical report (83.1% from the Gemini report), and the Flan-PaLM BBH figures (66.3% with CoT, not about 75%) (plus 21 minor fixes)
Version 10 · Nov 4, 2026, 10:33 PM
- Improve internal links (2026-10-13)
Version 8 · Oct 13, 2026, 10:15 AM
- Added 6 contextual internal links
Version 7 · Jul 23, 2026, 03:27 PM