MLLMs Know When Before Speaking: Revealing and Recovering Temporal Grounding via Attention Cues

arXiv:2605.21954 · cs.CV, cs.AI · Submitted 2026-05-21 · Read on arXiv

cs.CV, cs.AI

Submitted: 2026-05-21

Updated: 2026-09-27

Project page: https://ddz16.github.io/mllmsknowwhen.github.io

Terminology

• Grounding — 52× in this paper · explained in 3 episodes
• Aggregation — 12× in this paper · explained in 1 episode
• Benchmark — 11× in this paper · explained in 2 episodes
• Decode — 9× in this paper · explained in 2 episodes
• Adam — 6× in this paper · explained in 2 episodes
• Ablation — 5× in this paper · explained in 1 episode
• Accuracy — 4× in this paper · explained in 1 episode
• Calibration — 3× in this paper · explained in 7 episodes
• Fine-Tuning — 3× in this paper · explained in 5 episodes
• Grounding Contribution Score — 3× in this paper
• Reinforcement Learning — 3× in this paper · explained in 4 episodes
• Alignment — 2× in this paper · explained in 1 episode
• Core — 2× in this paper · explained in 2 episodes
• Pruning — 2× in this paper · explained in 4 episodes
• Active Regions — 1× in this paper · explained in 1 episode
• Entanglement — 1× in this paper · explained in 4 episodes
• Feature Extraction — 1× in this paper · explained in 2 episodes
• Multimodal Reasoning — 1× in this paper · explained in 5 episodes
• Robustness — 1× in this paper · explained in 5 episodes
• Transformer — 1× in this paper · explained in 2 episodes
• Two-Stage Framework — 1× in this paper · explained in 2 episodes
• Visual Grounding — 1× in this paper · explained in 2 episodes

Sources

• Qwen3-VL Technical Report
• Qwen2.5-VL Technical Report
• TEMPURA: Temporal Event Masked Prediction and Understanding for Reasoning in Action
• Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
• An Empirical Study on How Video-LLMs Answer Video Questions
• GPT-4o System Card
• Investigating The Functional Roles of Attention Heads in Vision Language Models: Evidence for Reasoning Modules
• See What You Are Told: Visual Attention Sink in Large Multimodal Models
• Interpreting Attention Heads for Image-to-Text Information Flow in Large Vision-Language Models
• VideoChat-Flash: Hierarchical Compression for Long-Context Video Modeling
• VideoChat-R1: Enhancing Spatio-Temporal Perception via Reinforcement Fine-Tuning
• Less Is More, but Where? Dynamic Token Compression via LLM-Guided Keyframe Prior
• TempSamp-R1: Effective Temporal Sampling with Reinforcement Fine-Tuning for Video LLMs
• Seeing but Not Believing: Probing the Disconnect Between Visual Attention and Answer Correctness in VLMs
• OpenAI GPT-5 System Card
• Grounded-VideoLLM: Sharpening Fine-grained Temporal Grounding in Video Large Language Models
• Time-R1: Post-Training Large Vision Language Model for Temporal Video Grounding
• MiMo-VL Technical Report
• Unveiling and Harnessing Hidden Attention Sinks: Enhancing Large Language Models without Training through Attention Calibration
• Tempo-R0: A Video-MLLM for Temporal Video Grounding through Efficient Temporal Sensing Reinforcement Learning

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