Decoding the Pulse of Reasoning VLMs in Multi-Image Understanding Tasks

TL;DR

PulseFocus is proposed, a training-free, inference-time method that structures CoT reasoning into interleaved plan/focus blocks with soft attention gating that sharpens attention focus and yields consistent improvements on multi-image benchmarks like BLINK benchmark and MuirBench.

Abstract

Multi-image reasoning remains a significant challenge for vision-language models (VLMs). We investigate a previously overlooked phenomenon: during chain-of-thought (CoT) generation, the text-to-image (T2I) attention of reasoning VLMs exhibits diffuse "pulses": sporadic and unfocused attention patterns that fail to concentrate on task-relevant images. We further reveal a systematic positional bias in attention allocation across images. Motivated by these observations, we propose PulseFocus, a training-free, inference-time method that structures CoT reasoning into interleaved plan/focus blocks with soft attention gating. By forcing the model to explicitly plan which image to examine and then gating decode-time attention to the referenced image, PulseFocus sharpens attention focus and yields consistent improvements on multi-image benchmarks like BLINK benchmark (+3.7%) and MuirBench (+1.07%).

Figures (5)

• Figure 1: Example case (from MuirBench). Baseline CoT fails to focus on the key evidence image (I5): token-level T2I colouring remains diffuse, and the model cannot recognize the second car. With PulseFocus, the <focus:I5> block becomes consistently image-aligned and the final answer is corrected from (C) to (B).
• Figure 2: Attention pulse visualization. T2I attention mass per image over CoT decode steps for a counting task (the same example as in Figure 1, with six input images). Top: baseline---attention is spread across images even when discussing a specific image. Bottom: with PulseFocus---attention concentrates on the image specified in the <focus:I> block. Colors indicate different images.
• Figure 3: Positional attention bias. Mean T2I attention mass per image position for InternVL3.5-8B on MuirBench. Earlier images receive disproportionately more attention regardless of task. Error bars show standard deviation across task types.
• Figure 4: PulseFocus overview. The model alternates between <plan> blocks (free attention, decides which image to examine next) and <focus:I> blocks (soft attention gate suppresses non-target images by $-\lambda$). Bottom: attention heatmaps contrasting standard decoding (left, diffuse) vs. gated decoding (right, concentrated).
• Figure 5: Image identity confusion (MuirBench # 359, 5 images).Left (Baseline): The model repeatedly examines "I2" but the token colours are dominantly red (I1) rather than blue (I2)---its verbal reference and actual visual attention are misaligned. It falsely concludes I2 matches the query, predicting (A). Right (PulseFocus): Each <focus:Ix> block's tokens correctly match the target image's colour, and the model concludes no match exists, predicting (B).