What Happens Before Decoding? Prefill Determines GUI Grounding in VLMs

• URL: http://arxiv.org/abs/2605.12549v1
• Date: Sun, 10 May 2026 07:04:07 GMT
• Title: What Happens Before Decoding? Prefill Determines GUI Grounding in VLMs
• Abstract summary: We study what happens during GUI grounding in Vision-Language Models (VLMs) and identify a previously overlooked bottleneck.<n>We show that grounding follows a two-stage paradigm: the prefill stage determines candidate UI elements, while the decoding stage subsequently refines the final coordinates.<n>We propose Re-Prefill, a training-free method that revisits inference by introducing an attention-guided second prefill stage to refine target selection.
• Score: 33.91859613266694
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Existing training-free approaches for GUI grounding often rely on multiple inference runs, such as iterative cropping or candidate aggregation, to identify target elements. Despite this additional computation, each forward pass still independently interprets the instruction and parses the visual layout, without enabling progressive interaction among visual tokens. In this paper, we study what happens during GUI grounding in Vision-Language Models (VLMs) and identify a previously overlooked bottleneck. We show that grounding follows a two-stage paradigm: the prefill stage determines candidate UI elements, while the decoding stage subsequently refines the final coordinates. This asymmetry establishes prefill as the critical step, as errors in candidate selection cannot be effectively corrected during decoding. Based on this observation, we propose Re-Prefill, a training-free method that revisits inference by introducing an attention-guided second prefill stage to refine target selection. Specifically, visual tokens that consistently receive high attention from the query position, i.e., the final token, across layers are extracted as a preliminary target hypothesis and appended to the input, together with the instruction hidden states, enabling the model to deeply re-think its decision before coordinate generation. Experiments across four VLMs and five benchmarks, including ScreenSpot-Pro, ScreenSpot-V2, OSWorld-G, UI-Vision, and MMBench-GUI, demonstrate consistent improvements without additional training, with gains of up to 4.3% on ScreenSpot-Pro. Code will be available at https://github.com/linjiaping1/Re-Prefill.

Related papers

• GUI-Lens: Coarse-to-Fine Cropping for GUI Grounding with General-Purpose VLMs [40.30152384879941]
  A vision-language model (VLM) may recognize a requested control without locating it precisely enough for interaction.<n>We introduce GUI-Lens, a coarse-to-fine grounding framework that allows a general-purpose VLM to determine the target through active visual observations.<n>Experiments show that GUI-Lens improves overall grounding accuracy by up to 24.9 percentage points and achieves state-of-the-art performance with GPT-5.5.
  arXiv  Detail & Related papers  (2026-08-04T07:47:37Z)
• [CLS] is Not Enough: Multi-Label Recognition via Patch-Level Inference and Adaptive Aggregation [20.637119409165418]
  PIAA formulates prediction as Patch-level Inference followed by Adaptive Aggregation.<n>We introduce an adaptive aggregation module that consolidates patch-level scores into the final multi-label prediction.<n>Experiments show that our method achieves strong improvements with minimal extra computation.
  arXiv  Detail & Related papers  (2026-05-25T13:19:12Z)
• Visual-Redundancy-Controlled Parallel Decoding for Diffusion-Based Multimodal Large Language Models [9.95016675032699]
  Diffusion-based multimodal large language models (dMLLMs) decode by iteratively predicting tokens at multiple masked positions in parallel.<n>Existing confidence-based decoding ranks masked positions independently and commits the top-K positions, largely ignoring whether the committed tokens provide complementary visual grounding.<n>We propose Visual-Redundancy-Controlled Decoding (VRCD), a training-free inference-time decoding method that uses token-to-image attention to prioritize visually complementary positions.
  arXiv  Detail & Related papers  (2026-05-25T13:16:51Z)
• See, Point, Refine: Multi-Turn Approach to GUI Grounding with Visual Feedback [8.080056970922483]
  Computer Use Agents (CUAs) fundamentally rely on graphical user interface (GUI) grounding to translate language instructions into executable screen actions.<n>Existing approaches typically rely on single-shot coordinate prediction, which lacks a mechanism for error correction and often fails in high-density interfaces.<n>In this technical report, we conduct an empirical study of pixel-precise cursor localization in coding environments.<n>We evaluate our approach across GPT-5.4, Claude, and Qwen on a suite of complex coding benchmarks, demonstrating that multi-turn refinement significantly outperforms state-of-the-art single-shot models in both click precision and
  arXiv  Detail & Related papers  (2026-04-14T17:55:46Z)
• HiViS: Hiding Visual Tokens from the Drafter for Speculative Decoding in Vision-Language Models [23.98782884568504]
  We propose Hiding Visual Tokens from the Drafter for Speculative Decoding in Vision-Language Models (HiViS)<n>HiViS is an explicit-implicit input decomposition framework that alleviates the inefficiency of Speculative Decoding in Vision-Language Models.<n>Our approach compresses the prefill sequence length of the drafter to only 0.7%-1.3% of the target VLM's input.
  arXiv  Detail & Related papers  (2025-09-28T15:05:21Z)
• GTA1: GUI Test-time Scaling Agent [97.58177633084915]
  Graphical user interface (GUI) agents autonomously complete tasks across platforms (eg, Linux) by sequentially decomposing user instructions into action proposals.<n>This paper investigates the aforementioned challenges with our textbfGUI textbfTest-time Scaling textbfAgent, namely GTA1.
  arXiv  Detail & Related papers  (2025-07-08T08:52:18Z)
• GUI-Actor: Coordinate-Free Visual Grounding for GUI Agents [93.49577107524176]
  We propose GUI-Actor, a VLM-based method for coordinate-free GUI grounding.<n>At its core, GUI-Actor introduces an attention-based action head that learns to align a dedicated ACTOR> token with all relevant visual patch tokens.<n>Experiments show that GUI-Actor outperforms prior state-of-the-art methods on multiple GUI action grounding benchmarks.
  arXiv  Detail & Related papers  (2025-06-03T17:59:08Z)
• LESS: Label-Efficient and Single-Stage Referring 3D Segmentation [55.06002976797879]
  Referring 3D is a visual-language task that segments all points of the specified object from a 3D point cloud described by a sentence of query. We propose a novel Referring 3D pipeline, Label-Efficient and Single-Stage, dubbed LESS, which is only under the supervision of efficient binary mask. We achieve state-of-the-art performance on ScanRefer dataset by surpassing the previous methods about 3.7% mIoU using only binary labels.
  arXiv  Detail & Related papers  (2024-10-17T07:47:41Z)
• SegVG: Transferring Object Bounding Box to Segmentation for Visual Grounding [56.079013202051094]
  We present SegVG, a novel method transfers the box-level annotation as signals to provide an additional pixel-level supervision for Visual Grounding. This approach allows us to iteratively exploit the annotation as signals for both box-level regression and pixel-level segmentation.
  arXiv  Detail & Related papers  (2024-07-03T15:30:45Z)
• Semantic Residual Prompts for Continual Learning [21.986800282078498]
  We show that our method significantly outperforms both state-of-the-art CL approaches and the zero-shot CLIP test. Our findings hold true even for datasets with a substantial domain gap w.r.t. the pre-training knowledge of the backbone model.
  arXiv  Detail & Related papers  (2024-03-11T16:23:38Z)
• Fine-grained Visual-Text Prompt-Driven Self-Training for Open-Vocabulary Object Detection [87.39089806069707]
  We propose a fine-grained Visual-Text Prompt-driven self-training paradigm for Open-Vocabulary Detection (VTP-OVD) During the adapting stage, we enable VLM to obtain fine-grained alignment by using learnable text prompts to resolve an auxiliary dense pixel-wise prediction task. Experiments show that our method achieves the state-of-the-art performance for open-vocabulary object detection, e.g., 31.5% mAP on unseen classes of COCO.
  arXiv  Detail & Related papers  (2022-11-02T03:38:02Z)
• Single-Stream Multi-Level Alignment for Vision-Language Pretraining [103.09776737512078]
  We propose a single stream model that aligns the modalities at multiple levels. We achieve this using two novel tasks: symmetric cross-modality reconstruction and a pseudo-labeled key word prediction. We demonstrate top performance on a set of Vision-Language downstream tasks such as zero-shot/fine-tuned image/text retrieval, referring expression, and VQA.
  arXiv  Detail & Related papers  (2022-03-27T21:16:10Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.