Revision History
LLaVA (Large Language and Vision Assistant) · 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.
- Correction: Corrected six wrong cells in the LLaVA-1.5 comparison table (InstructBLIP VQAv2/GQA/POPE, LLaVA-1.5 13B SEED-Bench, BLIP-2 POPE/SEED/LLaVA-Wild), the CLIP visual-token counts (256 at 224px, 576 at 336px), the ScienceQA 92.53% result (an ensemble with GPT-4), the misleading GPT-4o comparison for LLaVA-OneVision 72B, and the claim that LLaVA-NeXT was explicitly trained for Chinese; removed an unsupported $200 compute figure (plus 10 minor fixes)
Version 7 · Nov 4, 2026, 02:33 PM
- Improve internal links (2026-10-13)
Version 6 · Oct 13, 2026, 10:18 AM
- Added 6 contextual internal links
Version 5 · Jul 23, 2026, 03:28 PM