SOVABench: A Vehicle Surveillance Action Retrieval Benchmark for Multimodal Large Language Models

arXiv:2601.04824 · cs.CV · Submitted 2026-01-08 · Read on arXiv

cs.CV

Submitted: 2026-01-08

Updated: 2026-10-12

Code: https://github.com/oriol-rabasseda/sovabench

Terminology

• Benchmark — 6× in this paper · explained in 2 episodes
• Embeddings — 1× in this paper · explained in 2 episodes
• Training-Free Framework — 1× in this paper · explained in 2 episodes
• Vision-language models — 1× in this paper · explained in 2 episodes

Sources

• Task-aware Retrieval with Instructions
• Qwen2.5-VL Technical Report
• Think Then Embed: Generative Context Improves Multimodal Embedding
• VLM2Vec: Training Vision-Language Models for Massive Multimodal Embedding Tasks
• What's "up" with vision-language models? Investigating their struggle with spatial reasoning
• Exploring How Generative MLLMs Perceive More Than CLIP with the Same Vision Encoder
• VideoChat-R1: Enhancing Spatio-Temporal Perception via Reinforcement Fine-Tuning
• Towards General Text Embeddings with Multi-stage Contrastive Learning
• SurveillanceVQA-589K: A Benchmark for Comprehensive Surveillance Video-Language Understanding with Large Models
• TempCompass: Do Video LLMs Really Understand Videos?
• VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
• Beyond Captioning: Task-Specific Prompting for Improved VLM Performance in Mathematical Reasoning
• One Embedder, Any Task: Instruction-Finetuned Text Embeddings
• SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features
• EmbeddingGemma: Powerful and Lightweight Text Representations
• Multilingual E5 Text Embeddings: A Technical Report
• InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
• VideoCLIP: Contrastive Pre-training for Zero-shot Video-Text Understanding
• Qwen3 Technical Report
• MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training Recipe

Related papers

• Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
• AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
• Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
• MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
• TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
• A Survey on Efficient Vision-Language-Action Models