SVHalluc benchmark shows open-source audio-visual LLMs achieve near-random accuracy on semantic and temporal speech-vision alignment tasks while Gemini 2.5 Pro performs substantially better.
super hub Mixed citations
VideoLLaMA 2: Advancing Spatial-Temporal Modeling and Audio Understanding in Video-LLMs
Mixed citation behavior. Most common role is background (65%).
127
Pith papers citing it
Background
65% of classified citations
abstract
In this paper, we present the VideoLLaMA 2, a set of Video Large Language Models (Video-LLMs) designed to enhance spatial-temporal modeling and audio understanding in video and audio-oriented tasks. Building upon its predecessor, VideoLLaMA 2 incorporates a tailor-made Spatial-Temporal Convolution (STC) connector, which effectively captures the intricate spatial and temporal dynamics of video data. Additionally, we integrate an Audio Branch into the model through joint training, thereby enriching the multimodal understanding capabilities of the model by seamlessly incorporating audio cues. Comprehensive evaluations on multiple-choice video question answering (MC-VQA), open-ended video question answering (OE-VQA), and video captioning (VC) tasks demonstrate that VideoLLaMA 2 consistently achieves competitive results among open-source models and even gets close to some proprietary models on several benchmarks. Furthermore, VideoLLaMA 2 exhibits reasonable improvements in audio-only and audio-video question-answering (AQA & OE-AVQA) benchmarks over existing models. These advancements underline VideoLLaMA 2's superior performance in multimodal comprehension, setting a new standard for intelligent video analysis systems. All models are public to facilitate further research.
hub tools
citation-role summary
background 20
baseline 10
method 1
citation-polarity summary
claims ledger
- abstract In this paper, we present the VideoLLaMA 2, a set of Video Large Language Models (Video-LLMs) designed to enhance spatial-temporal modeling and audio understanding in video and audio-oriented tasks. Building upon its predecessor, VideoLLaMA 2 incorporates a tailor-made Spatial-Temporal Convolution (STC) connector, which effectively captures the intricate spatial and temporal dynamics of video data. Additionally, we integrate an Audio Branch into the model through joint training, thereby enriching the multimodal understanding capabilities of the model by seamlessly incorporating audio cues. Com
authors
co-cited works
representative citing papers
TraceAV-Bench is the first benchmark for multi-hop trajectory reasoning over long audio-visual videos, showing top models reach only 51-68% accuracy with substantial room for improvement.
VLMs hallucinate by prioritizing contradictory on-screen text over visual content, addressed via the VisualTextTrap benchmark with 6,057 human-validated samples and the VTHM-MoE dual-encoder framework using dimension-specific experts and adaptive routing.
EgoSound is a new benchmark with 7315 QA pairs across seven tasks to evaluate egocentric sound understanding in multimodal large language models.
ReQuest introduces an uncertainty-driven question-adaptive keyframe selector with rethinking routing and adaptive NMS that boosts long-form video QA accuracy on Video-MME, MLVU, and LongVideoBench without fine-tuning the base MLLM.
Introduces VidPair-Halluc benchmark of 1K background-controlled adversarial video pairs and 11K QA pairs generated via PairFlow pipeline to evaluate hallucination in LVMs.
MuseBench shows state-of-the-art MLLMs achieve only 48.29% accuracy on intent-level audiovisual arts understanding versus 87.18% for human experts.
Introduces VG-GUIBench benchmark and TASKER keyframe extraction algorithm that improves performance on VideoQA and video-guided agentic tasks.
AVLLMs route audio-visual information sequentially in video tasks and via parallel streams for interleaved items, allowing early token discard with little performance loss across models and scales.
Future-L1 interleaves latent visual spans with text in MLLM decoding, trained on a custom Future-L1-50K dataset via LA-DAPO RL, and reports SOTA gains on FutureBench (61.0 to 85.4) and TwiFF-Bench (2.44 to 3.04).
STORM teaches LVLMs to internalize spatial-temporal reasoning via bounded latent trajectories trained with generated thought videos in two stages, improving accuracy on VideoMME, MVBench and similar benchmarks while lowering inference overhead.
VideoOdyssey is a new benchmark featuring ultra-long videos (avg. 109 min) across 11 domains with multi-level continuous certificates (avg. 16 min for visual, 12.8 min for audio-visual) to diagnose MLLM limitations in continuous reasoning and omni-modal perception.
Video-LLMs exhibit directional motion blindness from a direction binding gap; DeltaDirect projector objective lifts synthetic accuracy to 85.4% and real accuracy by 21.9 points while preserving other video capabilities.
LatentOmni proposes a latent-space cross-modal reasoning framework that uses feature-level supervision and Omni-Sync Position Embedding to align and synchronize audio-visual latents, supported by a new 35K interleaved reasoning dataset and showing gains over text CoT baselines.
OmniPro is the first benchmark jointly evaluating omni-modal perception, proactive responding, and diverse streaming video understanding tasks using a dual-mode protocol on 2700 samples.
R3-Streaming uses cascaded control with age-aware memory forgetting and TB-GRPO reinforcement learning to reach SOTA scores of 57.92 on OVO-Bench and 76.36 on StreamingBench with 95-96% fewer visual tokens.
CoRDS selects a compact KV-cache subset via joint-space coreset coverage and log-det diversity to outperform token-wise heuristics on long-video VLM benchmarks.
ReTool-Video uses a 134-tool meta-augmented library and recursive grounding to translate abstract video intents into fine-grained multimodal operations, outperforming baselines on MVBench, MLVU, and Video-MME.
AdaFocus achieves better accuracy on long-video benchmarks with roughly 33 times fewer visual tokens by combining query-aware adaptive sampling and zero-cache disk-based refinement.
MMVIAD is the first multi-view continuous video dataset for industrial anomaly detection with four supported tasks, and the VISTA model improves average benchmark scores from 45.0 to 57.5 on unseen data while surpassing GPT-5.4.
TOC-Bench is a new diagnostic benchmark that reveals major weaknesses in temporal object consistency for Video-LLMs, including event counting, ordering, identity reasoning, and hallucination avoidance.
Temporal information in Video-LLMs is encoded well by video-centric encoders but disrupted by standard projectors; time-preserved MLPs plus AoT supervision yield 98.1% accuracy on arrow-of-time and gains on other temporal tasks.
A temperature-perturbed black-box attack infers video training membership in VideoLLMs with 0.68 AUC by exploiting sharper generation behavior on member samples.
GaLa uses hypergraph representations of objects and a TriView encoder with contrastive learning to improve vision-language models on procedural planning benchmarks.
citing papers explorer
-
SVHalluc: Benchmarking Speech-Vision Hallucination in Audio-Visual Large Language Models
SVHalluc benchmark shows open-source audio-visual LLMs achieve near-random accuracy on semantic and temporal speech-vision alignment tasks while Gemini 2.5 Pro performs substantially better.
-
TraceAV-Bench: Benchmarking Multi-Hop Trajectory Reasoning over Long Audio-Visual Videos
TraceAV-Bench is the first benchmark for multi-hop trajectory reasoning over long audio-visual videos, showing top models reach only 51-68% accuracy with substantial room for improvement.
-
When Text Hijacks Vision: Benchmarking and Mitigating Text Overlay-Induced Hallucination in Vision Language Models
VLMs hallucinate by prioritizing contradictory on-screen text over visual content, addressed via the VisualTextTrap benchmark with 6,057 human-validated samples and the VTHM-MoE dual-encoder framework using dimension-specific experts and adaptive routing.
-
EgoSound: Benchmarking Sound Understanding in Egocentric Videos
EgoSound is a new benchmark with 7315 QA pairs across seven tasks to evaluate egocentric sound understanding in multimodal large language models.
-
ReQuest: Rethinking-based Question-Aware Frame Selection for Long-Form Video QA
ReQuest introduces an uncertainty-driven question-adaptive keyframe selector with rethinking routing and adaptive NMS that boosts long-form video QA accuracy on Video-MME, MLVU, and LongVideoBench without fine-tuning the base MLLM.
-
No Place to Hide: Benchmarking Video Hallucination with Background-Controlled Pairs
Introduces VidPair-Halluc benchmark of 1K background-controlled adversarial video pairs and 11K QA pairs generated via PairFlow pipeline to evaluate hallucination in LVMs.
-
MuseBench: Benchmarking Intent-Level Audiovisual Arts Understanding in MLLMs
MuseBench shows state-of-the-art MLLMs achieve only 48.29% accuracy on intent-level audiovisual arts understanding versus 87.18% for human experts.
-
Bridging VideoQA and Video-Guided Agentic Tasks via Generalized Keyframe Extraction
Introduces VG-GUIBench benchmark and TASKER keyframe extraction algorithm that improves performance on VideoQA and video-guided agentic tasks.
-
From Senses to Decisions: The Information Flow of Auditory and Visual Perception in Multimodal LLMs
AVLLMs route audio-visual information sequentially in video tasks and via parallel streams for interleaved items, allowing early token discard with little performance loss across models and scales.
-
Imagine Before You Predict: Interleaved Latent Visual Reasoning for Video Event Prediction
Future-L1 interleaves latent visual spans with text in MLLM decoding, trained on a custom Future-L1-50K dataset via LA-DAPO RL, and reports SOTA gains on FutureBench (61.0 to 85.4) and TwiFF-Bench (2.44 to 3.04).
-
STORM: Internalized Modeling for Spatial-Temporal Reasoning in Video-Language Models
STORM teaches LVLMs to internalize spatial-temporal reasoning via bounded latent trajectories trained with generated thought videos in two stages, improving accuracy on VideoMME, MVBench and similar benchmarks while lowering inference overhead.
-
VideoOdyssey: A Benchmark for Ultra-Long-Context and Omni-Modal Video Understanding
VideoOdyssey is a new benchmark featuring ultra-long videos (avg. 109 min) across 11 domains with multi-level continuous certificates (avg. 16 min for visual, 12.8 min for audio-visual) to diagnose MLLM limitations in continuous reasoning and omni-modal perception.
-
Which Way Did It Move? Diagnosing and Overcoming Directional Motion Blindness in Video-LLMs
Video-LLMs exhibit directional motion blindness from a direction binding gap; DeltaDirect projector objective lifts synthetic accuracy to 85.4% and real accuracy by 21.9 points while preserving other video capabilities.
-
LatentOmni: Rethinking Omni-Modal Understanding via Unified Audio-Visual Latent Reasoning
LatentOmni proposes a latent-space cross-modal reasoning framework that uses feature-level supervision and Omni-Sync Position Embedding to align and synchronize audio-visual latents, supported by a new 35K interleaved reasoning dataset and showing gains over text CoT baselines.
-
OmniPro: A Comprehensive Benchmark for Omni-Proactive Streaming Video Understanding
OmniPro is the first benchmark jointly evaluating omni-modal perception, proactive responding, and diverse streaming video understanding tasks using a dual-mode protocol on 2700 samples.
-
An Efficient Streaming Video Understanding Framework with Agentic Control
R3-Streaming uses cascaded control with age-aware memory forgetting and TB-GRPO reinforcement learning to reach SOTA scores of 57.92 on OVO-Bench and 76.36 on StreamingBench with 95-96% fewer visual tokens.
-
CoRDS: Coreset-based Representative and Diverse Selection for Streaming Video Understanding
CoRDS selects a compact KV-cache subset via joint-space coreset coverage and log-det diversity to outperform token-wise heuristics on long-video VLM benchmarks.
-
ReTool-Video: Recursive Tool-Using Video Agents with Meta-Augmented Tool Grounding
ReTool-Video uses a 134-tool meta-augmented library and recursive grounding to translate abstract video intents into fine-grained multimodal operations, outperforming baselines on MVBench, MLVU, and Video-MME.
-
AdaFocus: Adaptive Relevance-Diversity Sampling with Zero-Cache Look-back for Efficient Long Video Understanding
AdaFocus achieves better accuracy on long-video benchmarks with roughly 33 times fewer visual tokens by combining query-aware adaptive sampling and zero-cache disk-based refinement.
-
MMVIAD: Multi-view Multi-task Video Understanding for Industrial Anomaly Detection
MMVIAD is the first multi-view continuous video dataset for industrial anomaly detection with four supported tasks, and the VISTA model improves average benchmark scores from 45.0 to 57.5 on unseen data while surpassing GPT-5.4.
-
TOC-Bench: A Temporal Object Consistency Benchmark for Video Large Language Models
TOC-Bench is a new diagnostic benchmark that reveals major weaknesses in temporal object consistency for Video-LLMs, including event counting, ordering, identity reasoning, and hallucination avoidance.
-
Tracing the Arrow of Time: Diagnosing Temporal Information Flow in Video-LLMs
Temporal information in Video-LLMs is encoded well by video-centric encoders but disrupted by standard projectors; time-preserved MLPs plus AoT supervision yield 98.1% accuracy on arrow-of-time and gains on other temporal tasks.
-
Membership Inference Attacks Against Video Large Language Models
A temperature-perturbed black-box attack infers video training membership in VideoLLMs with 0.68 AUC by exploiting sharper generation behavior on member samples.
-
GaLa: Hypergraph-Guided Visual Language Models for Procedural Planning
GaLa uses hypergraph representations of objects and a TriView encoder with contrastive learning to improve vision-language models on procedural planning benchmarks.
-
Watching Movies Like a Human: Egocentric Emotion Understanding for Embodied Companions
Creates the first egocentric screen-view movie emotion benchmark and demonstrates that cinematic models drop sharply in Macro-F1 on realistic robot-like viewing conditions while domain-specific training improves robustness.
-
Chain of Modality: From Static Fusion to Dynamic Orchestration in Omni-MLLMs
Chain of Modality dynamically orchestrates multimodal input topologies and bifurcates cognitive execution to overcome static fusion biases in Omni-MLLMs.
-
Don't Let the Video Speak: Audio-Contrastive Preference Optimization for Audio-Visual Language Models
Audio-Contrastive Preference Optimization (ACPO) mitigates audio hallucination in AVLMs via output-contrastive and input-contrastive objectives that enforce faithful audio grounding.
-
BoxTuning: Directly Injecting the Object Box for Multimodal Model Fine-Tuning
By drawing object boxes and motion trails visually on video frames instead of serializing coordinates as text, BoxTuning reduces token costs dramatically and improves accuracy on video question answering benchmarks.
-
AdaSpark: Adaptive Sparsity for Efficient Long-Video Understanding
AdaSpark delivers up to 57% FLOP reduction in Video-LLMs for long videos through adaptive cube- and token-level sparsity without apparent loss in performance on hour-scale benchmarks.
-
Bridging Time and Space: Decoupled Spatio-Temporal Alignment for Video Grounding
Bridge-STG decouples spatio-temporal alignment via semantic bridging and query-guided localization modules to achieve state-of-the-art m_vIoU of 34.3 on VidSTG among MLLM methods.
-
SVAgent: Storyline-Guided Long Video Understanding via Cross-Modal Multi-Agent Collaboration
SVAgent improves long video question answering by constructing storylines via multi-agent collaboration and aligning cross-modal predictions for more robust, human-like reasoning.
-
From Verbatim to Gist: Distilling Pyramidal Multimodal Memory via Semantic Information Bottleneck for Long-Horizon Video Agents
MM-Mem distills video input through a hierarchical memory of sensory buffer, episodic stream, and symbolic schema, optimized by a semantic information bottleneck and SIB-GRPO, to achieve SOTA on long-horizon video benchmarks.
-
VideoSearch-R1: Iterative Video Retrieval and Reasoning via Soft Query Refinement
VideoSearch-R1 achieves SOTA on VCMR across three datasets via iterative retrieval, latent-space soft query refinement, and GRPO training.
-
VisReflect: Latent Visual Reflection for Fine-Grained Perception in Long Visual Context
VisReflect generates continuous latent visual reflections to emphasize relevant visual features and guide attention in LVLMs, yielding 4.1% gains on image benchmarks and 1.8% on video benchmarks with 44% less inference time than zooming methods.
-
HAT-4D: Lifting Monocular Video for 4D Multi-Object Interactions via Human-Agent Collaboration
HAT-4D presents an agentic VLM-plus-human-in-the-loop pipeline for monocular 4D multi-object interaction reconstruction and releases the MVOIK-4D benchmark.
-
Towards Accurate and Robust Surveillance Roadside IVD via Trackletized Audio-Visual Reasoning
TAVR-IVD uses multi-object tracking to create per-vehicle tracklets for audio-visual idling detection, claiming higher SNR, temporal stability, explicit spatial alignment, and better domain adaptation than full-frame fusion.
-
HPP: Hierarchical Programmatic Probing for Long Video Understanding by Decoupling Perception and Reasoning
HPP decouples perception from reasoning in long-video VLMs by having an LLM run iterative programmatic probes on hierarchically segmented video, reporting gains on LongVideoBench, EgoSchema, VideoMME, and MLVU.
-
CARE: Competence-Aware Reward Shaping for Adaptive Reasoning Length in Video-MLLMs
CARE uses exponential moving average competence estimates to progressively shift RL rewards from exploration-oriented long reasoning to efficiency-oriented concise reasoning in video-MLLMs, with batch normalization and posterior amplification, yielding accuracy gains and shorter traces.
-
Reasoning as Intersection: Consensus-Frame Alignment for Visual Focus in Video-MLLMs
CF-GRPO creates a consensus frame prior from intrinsic video cues and aligns it with model frame-use scores via a reward signal to enable evidence-aware reasoning in Video-MLLMs without temporal annotations.
-
HYDRA-X: Native Unified Multimodal Models with Holistic Visual Tokenizers
HYDRA-X presents the first unified multimodal model using a single ViT for holistic image-video tokenization, with ablations on attention and compression plus a latent-level editing improvement.
-
Rethinking RAG in Long Videos: What to Retrieve and How to Use It?
Introduces V-RAGBench benchmark and CARVE method that selects per-chunk retrieval configurations via parallel retrievers and adaptive reranking, outperforming eight VideoRAG baselines.
-
Don't Pause: Streaming Video-Language Synchrony for Online Video Understanding
LyraV uses FDTC and SToP for per-frame incremental decoding to reach 98.29% video synchrony at 3.89 FPS while preserving general understanding.
-
StoryVideoQA: Scaling Deep Video Understanding with a Large-Scale, Multi-Genre and Auto-Generated Dataset
StoryVideoQA provides the largest auto-generated deep video understanding dataset to date with 363K QAs across TV and movies, paired with the PlotTree agent for hierarchical plot-based reasoning that existing VideoQA models struggle to match.
-
VTI-CoT: Visual-Textual Interleaved Chain of Thought for Video Reasoning
VTI-CoT proposes a visual-textual interleaved chain-of-thought method for video reasoning, built via automated annotation and OCR compression, claiming SOTA performance and better training efficiency on same-scale models.
-
GOPAgen: Motion-Aware and Efficient Agentic Long-Video Understanding with Structural Memory and Hierarchical Reasoning
GOPAgen proposes integrating video codec GOPs with a motion agent, GOP tree reasoning, structural memory, and motion vector database to improve efficiency and motion detail in agentic long-video VQA, reporting gains on MotionBench and EgoSchema.
-
InfoMerge: Information-aware Token Compression for Efficient Video Large Language Models
InfoMerge proposes a training-free visual token compression method for Video-LLMs that uses Temporal Fingerprint Difference for redundancy estimation and Content-Aware Budget Allocation to retain 98.8% performance with 85% fewer tokens.
-
MTAVG-Bench 2.0: Diagnosing Failure Modes of Cinematic Expressiveness in Multi-Talker Audio-Video Generation
MTAVG-Bench 2.0 is a new benchmark that evaluates omni LLMs on diagnosing high-level cinematic failures in multi-talker audio-video generation using a taxonomy of acting, narrative, atmosphere, and audio-visual language.
-
O-MARC: Omni Memory-Augmented Compression Distillation for Efficient Video Understanding
O-MARC is a compression distillation framework that lets compact omnimodal models maintain or exceed full-token performance on video QA while cutting latency and memory by about 35%.
-
Not All Modalities Are Equal: Instruction-Aware Gating for Multimodal Videos
UniMVU applies instruction-conditioned inner-modality and modality-level gates to adaptively fuse multiple video modalities, achieving gains of up to 13.5 CIDEr on six benchmarks including AVQA and MVBench.
-
Enhancing Visual Token Representations for Video Large Language Models via Training-Free Spatial-Temporal Pooling and Gridding
ST-GridPool improves video LLM performance via hierarchical temporal gridding and norm-based spatial pooling on visual tokens without training.