DoubleTake targets focused verification of visual evidence in long-document QA
A new arXiv paper, DoubleTake, addresses vision-language models answering long-document questions incorrectly even when retrieved pages contain sufficient evidence, proposing focused rereading/verification of visual evidence.
Read the full story at arXiv cs.CV ↗
Timeline · 1 report
- 2026-11-16 05:00 · arXiv cs.CV
DoubleTake: Focused Verification of Visual Evidence for Long-Document Understanding