# Recurrent Attention-based Token Selection for Efficient Streaming Video-LLMs

arXiv: https://arxiv.org/abs/2510.17364 · HTML: https://mlsanity.com/paper/2510.17364/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 33 of its 40 references that are in the map.
- First posted on arXiv 20 Oct 2025.

## Builds on (most cited first, 30 of 33)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation](https://mlsanity.com/paper/2201.12086/) (2022)
- [Qwen Technical Report](https://mlsanity.com/paper/2309.16609/) (2023)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://mlsanity.com/paper/2203.12602/) (2022)
- [Video-ChatGPT: Towards Detailed Video Understanding via Large Vision and Language Models](https://mlsanity.com/paper/2306.05424/) (2023)
- [NExT-QA:Next Phase of Question-Answering to Explaining Temporal Actions](https://mlsanity.com/paper/2105.08276/) (2021)
- [Multimodal Few-Shot Learning with Frozen Language Models](https://mlsanity.com/paper/2106.13884/) (2021)
- [LLaMA-VID: An Image is Worth 2 Tokens in Large Language Models](https://mlsanity.com/paper/2311.17043/) (2023)
- [MovieChat: From Dense Token to Sparse Memory for Long Video Understanding](https://mlsanity.com/paper/2307.16449/) (2023)
- [VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking](https://mlsanity.com/paper/2303.16727/) (2023)
- [Less is More: ClipBERT for Video-and-Language Learning via Sparse Sampling](https://mlsanity.com/paper/2102.06183/) (2021)
- [InternVideo: General Video Foundation Models via Generative and Discriminative Learning](https://mlsanity.com/paper/2212.03191/) (2022)
- [LVBench: An Extreme Long Video Understanding Benchmark](https://mlsanity.com/paper/2406.08035/) (2024)
- [Chat-UniVi: Unified Visual Representation Empowers Large Language Models with Image and Video Understanding](https://mlsanity.com/paper/2311.08046/) (2023)
- [MLVU: Benchmarking Multi-task Long Video Understanding](https://mlsanity.com/paper/2406.04264/) (2024)
- [MA-LMM: Memory-Augmented Large Multimodal Model for Long-Term Video Understanding](https://mlsanity.com/paper/2404.05726/) (2024)
- [Video-XL: Extra-Long Vision Language Model for Hour-Scale Video Understanding](https://mlsanity.com/paper/2409.14485/) (2024)
- [In-context Autoencoder for Context Compression in a Large Language Model](https://mlsanity.com/paper/2307.06945/) (2023)
- [Flash-VStream: Memory-Based Real-Time Understanding for Long Video Streams](https://mlsanity.com/paper/2406.08085/) (2024)
- [Streaming Long Video Understanding with Large Language Models](https://mlsanity.com/paper/2405.16009/) (2024)
- [Tarsier: Recipes for Training and Evaluating Large Video Description Models](https://mlsanity.com/paper/2407.00634/) (2024)
- [Fit and Prune: Fast and Training-free Visual Token Pruning for Multi-modal Large Language Models](https://mlsanity.com/paper/2409.10197/) (2024)
- [Streaming Video Question-Answering with In-context Video KV-Cache Retrieval](https://mlsanity.com/paper/2503.00540/) (2025)
- [CG-Bench: Clue-grounded Question Answering Benchmark for Long Video Understanding](https://mlsanity.com/paper/2412.12075/) (2024)
- [Simple Unsupervised Keyphrase Extraction using Sentence Embeddings](https://mlsanity.com/paper/1801.04470/) (2018)
- [Sports-QA: A Large-Scale Video Question Answering Benchmark for Complex and Professional Sports](https://mlsanity.com/paper/2401.01505/) (2024)
- [Goldfish: Vision-Language Understanding of Arbitrarily Long Videos](https://mlsanity.com/paper/2407.12679/) (2024)

## Built on by (most cited first, 9 of 9)

- [Video Streaming Thinking: VideoLLMs Can Watch and Think Simultaneously](https://mlsanity.com/paper/2603.12262/) (2026)
- [StreamMeCo: Long-Term Agent Memory Compression for Efficient Streaming Video Understanding](https://mlsanity.com/paper/2604.09000/) (2026)
- [Position: Modular Memory is the Key to Continual Learning Agents](https://mlsanity.com/paper/2603.01761/) (2026)
- [MemoryCard: Topic-Aware Multi-Modal Clue Compression for Long-Video Question Answering](https://mlsanity.com/paper/2606.05917/) (2026)
- [Going Down Memory Lane: Scaling Tokens for Video Stream Understanding with Dynamic KV-Cache Memory](https://mlsanity.com/paper/2602.18434/) (2026)
- [ViCoStream: Streaming VideoLLMs Can Run Beyond 100 FPS with Stage-Wise Coordinated Inference](https://mlsanity.com/paper/2606.19849/) (2026)
- [Beyond Retrieval: Progressive Latent Memory Evolution for Streaming Video Understanding](https://mlsanity.com/paper/2609.04131/) (2026)
- [Ego: Embedding-Guided Personalization of Vision-Language Models](https://mlsanity.com/paper/2603.09771/) (2026)
- [MemForest: Efficient Agent Memory Management via EventTree Partitioning and Progressive Merging](https://mlsanity.com/paper/2609.08273/) (2026)

## Abstract (arXiv)

Video Large Language Models (Video-LLMs) excel at understanding videos in-context, provided they have full access to the video when answering queries. However, these models face challenges in streaming scenarios where hour-long videos must be processed online, and questions need timely responses. In this work, we propose a training-free approach compatible with standard Video-LLMs, leveraging three key concepts: 1) LLM-informed selection of visual tokens to identify those that the LLM has attended to and contributed to its understanding of each short clip. Our attention-based selection allows us to discard up to ~95% of unimportant visual tokens with minimal performance loss; 2) Recurrent processing of past selected tokens to generate temporally coherent understanding of each processed clip; 3) Caption-based question answering for lightweight and accurate responses. Our method achieves state-of-the-art performance on streaming video benchmarks, striking a balance between efficiency and effectiveness.