VGenST-Bench is a new video benchmark for MLLM spatio-temporal reasoning built via generative synthesis, a multi-agent pipeline with human oversight, a 3x2x2 taxonomy, and hierarchical tasks separating perception from reasoning.
Eyes wide shut? exploring the visual shortcomings of multimodal llms
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
citation-role summary
background 2
dataset 1
citation-polarity summary
years
2026 7representative citing papers
Introduces Synergistic Faithfulness metric based on Shapley Interaction Index to evaluate cross-modal synergy in VLM explainers, revealing over-reliance on visual salience in existing methods.
Vision-OPD transfers an MLLM's privileged regional perception to its full-image policy through on-policy token-level self-distillation, yielding competitive results on fine-grained visual benchmarks.
SpatialForge synthesizes 10 million spatial QA pairs from in-the-wild 2D images to train VLMs for better depth ordering, layout, and viewpoint-dependent reasoning.
RIS improves MLLM latent visual reasoning by retrieving spatial-semantic evidence, integrating it via attention bottlenecks, and synthesizing it with language transition tokens, yielding gains on V*, HRBench, MMVP, and BLINK benchmarks.
EAGLE achieves up to 94.4% anomaly detection accuracy on MVTec-AD and 88.1% on VisA by guiding frozen MLLMs with expert-derived thresholds and confidence-aware attention without parameter updates.
Textual prompts override visual cues in a medical MLLM, reducing accuracy from 75% to 46% on a hemorrhage versus drusen task even when visual grounding is present.
citing papers explorer
-
VGenST-Bench: A Benchmark for Spatio-Temporal Reasoning via Active Video Synthesis
VGenST-Bench is a new video benchmark for MLLM spatio-temporal reasoning built via generative synthesis, a multi-agent pipeline with human oversight, a 3x2x2 taxonomy, and hierarchical tasks separating perception from reasoning.
-
Measuring Cross-Modal Synergy: A Benchmark for VLM Explainability
Introduces Synergistic Faithfulness metric based on Shapley Interaction Index to evaluate cross-modal synergy in VLM explainers, revealing over-reliance on visual salience in existing methods.
-
Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Policy Self-Distillation
Vision-OPD transfers an MLLM's privileged regional perception to its full-image policy through on-policy token-level self-distillation, yielding competitive results on fine-grained visual benchmarks.
-
SpatialForge: Bootstrapping 3D-Aware Spatial Reasoning from Open-World 2D Images
SpatialForge synthesizes 10 million spatial QA pairs from in-the-wild 2D images to train VLMs for better depth ordering, layout, and viewpoint-dependent reasoning.
-
Retrieve, Integrate, and Synthesize: Spatial-Semantic Grounded Latent Visual Reasoning
RIS improves MLLM latent visual reasoning by retrieving spatial-semantic evidence, integrating it via attention bottlenecks, and synthesizing it with language transition tokens, yielding gains on V*, HRBench, MMVP, and BLINK benchmarks.
-
EAGLE: Expert-Augmented Attention Guidance for Tuning-Free Industrial Anomaly Detection in Multimodal Large Language Models
EAGLE achieves up to 94.4% anomaly detection accuracy on MVTec-AD and 88.1% on VisA by guiding frozen MLLMs with expert-derived thresholds and confidence-aware attention without parameter updates.
-
When Prompts Mislead: Textual Dominance and Diagnostic Bias in MLLMs
Textual prompts override visual cues in a medical MLLM, reducing accuracy from 75% to 46% on a hemorrhage versus drusen task even when visual grounding is present.