# DeltaS: Reading the Gated Linear Attention State for KV Cache Eviction in Streaming Video

arXiv: https://arxiv.org/abs/2609.27470 · HTML: https://mlsanity.com/paper/2609.27470/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Fast autoregressive visual generation, one of 12 directions in Efficient LLM inference and serving.
- Builds on 36 of its 41 references that are in the map.
- First posted on arXiv 23 Sep 2026.

## Placed in

- Efficient LLM inference and serving / [Fast autoregressive visual generation](https://mlsanity.com/inference/fast-autoregressive-visual-generation/)

## Builds on (most cited first, 30 of 36)

- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis](https://mlsanity.com/paper/2405.21075/) (2024)
- [SnapKV: LLM Knows What You are Looking for Before Generation](https://mlsanity.com/paper/2404.14469/) (2024)
- [EgoSchema: A Diagnostic Benchmark for Very Long-form Video Language Understanding](https://mlsanity.com/paper/2308.09126/) (2023)
- [LongVideoBench: A Benchmark for Long-context Interleaved Video-Language Understanding](https://mlsanity.com/paper/2407.15754/) (2024)
- [KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache](https://mlsanity.com/paper/2402.02750/) (2024)
- [Gated Delta Networks: Improving Mamba2 with Delta Rule](https://mlsanity.com/paper/2412.06464/) (2024)
- [Linear Transformers Are Secretly Fast Weight Programmers](https://mlsanity.com/paper/2102.11174/) (2021)
- [LVBench: An Extreme Long Video Understanding Benchmark](https://mlsanity.com/paper/2406.08035/) (2024)
- [Titans: Learning to Memorize at Test Time](https://mlsanity.com/paper/2501.00663/) (2024)
- [MLVU: Benchmarking Multi-task Long Video Understanding](https://mlsanity.com/paper/2406.04264/) (2024)
- [Kimi Linear: An Expressive, Efficient Attention Architecture](https://mlsanity.com/paper/2510.26692/) (2025)
- [MiniMax-01: Scaling Foundation Models with Lightning Attention](https://mlsanity.com/paper/2501.08313/) (2025)
- [StreamingBench: Assessing the Gap for MLLMs to Achieve Streaming Video Understanding](https://mlsanity.com/paper/2411.03628/) (2024)
- [OVO-Bench: How Far is Your Video-LLMs from Real-World Online Video Understanding?](https://mlsanity.com/paper/2501.05510/) (2025)
- [Streaming Video Question-Answering with In-context Video KV-Cache Retrieval](https://mlsanity.com/paper/2503.00540/) (2025)
- [StreamingVLM: Real-Time Understanding for Infinite Video Streams](https://mlsanity.com/paper/2510.09608/) (2025)
- [A Simple and Effective L2 Norm-Based Strategy for KV Cache Compression](https://mlsanity.com/paper/2406.11430/) (2024)
- [StreamForest: Efficient Online Video Understanding with Persistent Event Memory](https://mlsanity.com/paper/2509.24871/) (2025)
- [InfiniPot-V: Memory-Constrained KV Cache Compression for Streaming Video Understanding](https://mlsanity.com/paper/2506.15745/) (2025)
- [StreamMem: Query-Agnostic KV Cache Memory for Streaming Video Understanding](https://mlsanity.com/paper/2508.15717/) (2025)
- [Video-MME-v2: Towards the Next Stage in Benchmarks for Comprehensive Video Understanding](https://mlsanity.com/paper/2604.05015/) (2026)
- [LiveVLM: Efficient Online Video Understanding via Streaming-Oriented KV Cache and Retrieval](https://mlsanity.com/paper/2505.15269/) (2025)
- [KeyDiff: Key Similarity-Based KV Cache Eviction for Long-Context LLM Inference in Resource-Constrained Environments](https://mlsanity.com/paper/2504.15364/) (2025)
- [StreamingTOM: Streaming Token Compression for Efficient Video Understanding](https://mlsanity.com/paper/2510.18269/) (2025)
- [HERMES: KV Cache as Hierarchical Memory for Efficient Streaming Video Understanding](https://mlsanity.com/paper/2601.14724/) (2026)
- [A Simple Baseline for Streaming Video Understanding](https://mlsanity.com/paper/2604.02317/) (2026)
- [StreamKV: Streaming Video Question-Answering with Segment-based KV Cache Retrieval and Compression](https://mlsanity.com/paper/2511.07278/) (2025)
- [Video Streaming Thinking: VideoLLMs Can Watch and Think Simultaneously](https://mlsanity.com/paper/2603.12262/) (2026)
- [The Pitfalls of KV Cache Compression](https://mlsanity.com/paper/2510.00231/) (2025)

## Abstract (arXiv)

Recent video-language models increasingly adopt hybrid architectures that interleave linear and full attention layers for efficient long-context processing. While the recurrent state of linear attention remains fixed in size, the KV cache of full attention continues to grow with the video stream, making eviction necessary under a bounded memory budget. The key challenge in streaming is that eviction must occur before the question arrives, so what to retain has to be decided without the question. Existing eviction methods derive token scores from the KV cache itself, using position, attention, or key-value representations, and attention-based scores further require proxy queries or extra computation. Hybrid backbones offer another source of signal. In gated-delta linear attention, the recurrent state is updated by the residual between each input and what can already be retrieved from the state, so its change over a chunk of frames reflects how much new information the chunk brings. We propose DeltaS, a query-agnostic, training-free method that retains video chunks inducing larger normalized state change, or state drift. In a controlled comparison with the budget and retention policy held fixed, state drift outperforms position-, attention-, and key-value-based signals. With a signal costing only 1.9% of the forward pass, DeltaS surpasses the strongest query-agnostic bounded-memory baseline by 2.1 points on average across six long-video benchmarks and by 5.6 points on the longest benchmark. These results suggest that the two memories of hybrid architectures can work cooperatively. Code is available at https://github.com/MaumAI-Company/DeltaS.