Metraipolis › archive › dc_76ff25e81b99

TestHallVQA: Exploring LVLMs' Document-Level Reasoning under Redundant Contexts from Scientific Exams — arXiv:2609.13158

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.13158
License (as archived)
CC0-1.0
Archived
2026-10-15T23:13:24.617Z
SHA-256 of the archived text
8f56a0f1787b29c71ed085350b666a3b042115a76634dc06aed6e4531b69153f
Stable URL of this page
https://metraipolis.com/documento/dc_76ff25e81b99

How to cite

TestHallVQA: Exploring LVLMs' Document-Level Reasoning under Redundant Contexts from Scientific Exams — arXiv:2609.13158 — arXiv (en), https://arxiv.org/abs/2609.13158 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_76ff25e81b99 (Metraipolis archive dc_76ff25e81b99, sha256 8f56a0f1787b29c71ed085350b666a3b042115a76634dc06aed6e4531b69153f)
BibTeX
@misc{dc76ff25e81b99,
  title = {TestHallVQA: Exploring LVLMs' Document-Level Reasoning under Redundant Contexts from Scientific Exams — arXiv:2609.13158},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.13158},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_76ff25e81b99; sha256 8f56a0f1787b29c71ed085350b666a3b042115a76634dc06aed6e4531b69153f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.13158. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large Vision--Language Models (LVLMs) are increasingly expected to perform visual question answering (VQA) over planar media. However, existing planar VQA benchmarks typically emphasize isolated challenges: some emphasize long-document understanding with limited reasoning depth, while others require complex visual reasoning but remain restricted to single-page, noise-free settings. Moreover, through theoretical analysis, we identify the impact of irrelevant visual tokens, which leads to measurable performance degradation but has received little attention with respect to systematic quantification. To address these limitations, we introduce TestHallVQA, a multi-image VQA benchmark that simultaneously embodies document-level scale and the difficulty of human examinations, while providing comprehensive task coverage. Leveraging TestHallVQA's ability to controllably inject multi-level contextual redundancy, we further propose a novel metric, F1-R\textsuperscript{2}, which jointly quantifies LVLMs' computational reasoning capability and their evidence retrieval robustness against document-level redundancy. Extensive experiments and analyses on mainstream LVLMs reveal their latent deficiencies across multiple dimensions, offering concrete insights and directions for future research. The associated datasets, code, and complete theoretical derivations are available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-15 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.13158 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-15T22:51:09.813Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-15?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).