mlsanityData to 14 Oct 2026

Towards Reliable and Interpretable Document Question Answering via VLMs

arXiv 2509.10129 ↗First posted 12 Sep 2025cs.CL cs.IRcited by 2

On the map: Multimodal LLMs / vision-language models → Document understanding and multimodal RAG

Abstract (arXiv)

Vision-Language Models (VLMs) have shown strong capabilities in document understanding, particularly in identifying and extracting textual information from complex documents. Despite this, accurately localizing answers within documents remains a major challenge, limiting both interpretability and real-world applicability. To address this, we introduce DocExplainerV0, a plug-and-play bounding-box prediction module that decouples answer generation from spatial localization. This design makes it applicable to existing VLMs, including proprietary systems where fine-tuning is not feasible. Through systematic evaluation, we provide quantitative insights into the gap between textual accuracy and spatial grounding, showing that correct answers often lack reliable localization. Our standardized framework highlights these shortcomings and establishes a benchmark for future research toward more interpretable and robust document information extraction VLMs.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in Document understanding and multimodal RAG, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 29 of its 35 references that are in the map. First posted on arXiv 12 Sep 2025.

Builds on

29 of 35 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025SmolDocling: An ultra-compact vision-language model for end-to-end multi-modal document conversion14 Mar
  2. 2024DocVLM: Make Your VLM an Efficient Reader11 Dec

Other references · 27

  1. 2019Towards VQA Models That Can Readcited by 1,718
  2. 2022ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoningcited by 1,479
  3. 2020DocVQA: A Dataset for VQA on Document Imagescited by 1,317
  4. 2025SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Featurescited by 1,188
  5. 2023Vision-Language Models for Vision Tasks: A Surveycited by 839
  6. 2022LayoutLMv3: Pre-training for Document AI with Unified Text and Image Maskingcited by 485
  7. 2021OCR-free Document Understanding Transformercited by 407
  8. 2022Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understandingcited by 366
  9. 2024Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Modelscited by 329
  10. 2025SmolVLM: Redefining small and efficient multimodal modelscited by 237
  11. 2024Building and better understanding vision-language models: insights and future directionscited by 173
  12. 2023Document Understanding Dataset and Evaluation (DUDE)cited by 135
15 more
  1. 2023DocLLM: A layout-aware generative language model for multimodal document understandingcited by 116
  2. 2024A Bounding Box is Worth One Token: Interleaving Layout and Text in a Large Language Model for Document Understandingcited by 57
  3. 2023DocFormerv2: Local Features for Document Understandingcited by 46
  4. 2023LMDX: Language Model-based Document Information Extraction and Localizationcited by 37
  5. 2024From Training-Free to Adaptive: Empirical Insights into MLLMs' Understanding of Detection Informationcited by 16
  6. 2025How Well Does GPT-4o Understand Vision? Evaluating Multimodal Foundation Models on Standard Computer Vision Taskscited by 14
  7. 2024ANLS* -- A Universal Document Processing Metric for Generative Large Language Modelscited by 10
  8. 2025Attention, Please! PixelSHAP Reveals What Vision-Language Models Actually Focus Oncited by 9
  9. 2024DLaVA: Document Language and Vision Assistant for Answer Localization with Enhanced Interpretability and Trustworthinesscited by 8
  10. 2025BoundingDocs: a Unified Dataset for Document Question Answering with Spatial Annotationscited by 8
  11. 2025DocVXQA: Context-Aware Visual Explanations for Document Question Answeringcited by 5
  12. 2024DocXplain: A Novel Model-Agnostic Explainability Method for Document Image Classificationcited by 4
  13. 2025Where is this coming from? Making groundedness count in the evaluation of Document VQA modelscited by 4
  14. 2024LLM-wrapper: Black-Box Semantic-Aware Adaptation of Vision-Language Models for Referring Expression Comprehensioncited by 3
  15. 2025Evaluating Robustness of LLMs in Question Answering on Multilingual Noisy OCR Datacited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Comparative Study of Domain-adapted VLMs for General Document Visual Question Answeringcited by 3
  2. 2026LED: A Benchmark for Evaluating Layout Error Detection in Document Analysiscited by 1

Similar, not cited

Close in wording, no citation either way

  1. 2025ARIAL: An Agentic Framework for Document VQA with Precise Answer Localization
  2. 2026Towards Self-Explainable Document Visual Question Answering with Chain-of-Explanation Predictions
  3. 2025Spatially Grounded Explanations in Vision Language Models for Document Visual Question Answering
  4. 2025DocVAL: Validated Chain-of-Thought Distillation for Grounded Document VQA
  5. 2025MGA-VQA: Secure and Interpretable Graph-Augmented Visual Question Answering with Memory-Guided Protection Against Unauthorized Knowledge Use
  6. 2026DocTrace: Towards Traceable Long Document VQA via Hierarchical Evidence Graph Reasoning
  7. 2025DRISHTIKON: Visual Grounding at Multiple Granularities in Documents
  8. 2025Joint Extraction Matters: Prompt-Based Visual Question Answering for Multi-Field Document Information Extraction