# RefreshKV: Updating Small KV Cache During Long-form Generation

arXiv: https://arxiv.org/abs/2411.05787 · HTML: https://mlsanity.com/paper/2411.05787/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Placed in KV cache optimization, one of 12 directions in Efficient LLM inference and serving.
- Builds on 26 of its 32 references that are in the map.
- First posted on arXiv 8 Nov 2024.

## Placed in

- Efficient LLM inference and serving / [KV cache optimization](https://mlsanity.com/inference/kv-cache-optimization/)

## Builds on (most cited first, 26 of 26)

- [Longformer: The Long-Document Transformer](https://mlsanity.com/paper/2004.05150/) (2020)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [Big Bird: Transformers for Longer Sequences](https://mlsanity.com/paper/2007.14062/) (2020)
- [Generating Long Sequences with Sparse Transformers](https://mlsanity.com/paper/1904.10509/) (2019)
- [LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding](https://mlsanity.com/paper/2308.14508/) (2023)
- [GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints](https://mlsanity.com/paper/2305.13245/) (2023)
- [RULER: What's the Real Context Size of Your Long-Context Language Models?](https://mlsanity.com/paper/2404.06654/) (2024)
- [SnapKV: LLM Knows What You are Looking for Before Generation](https://mlsanity.com/paper/2404.14469/) (2024)
- [PyTorch FSDP: Experiences on Scaling Fully Sharded Data Parallel](https://mlsanity.com/paper/2304.11277/) (2023)
- [KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache](https://mlsanity.com/paper/2402.02750/) (2024)
- [KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantization](https://mlsanity.com/paper/2401.18079/) (2024)
- [Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference](https://mlsanity.com/paper/2406.10774/) (2024)
- [Model Tells You What to Discard: Adaptive KV Cache Compression for LLMs](https://mlsanity.com/paper/2310.01801/) (2023)
- [8-bit Optimizers via Block-wise Quantization](https://mlsanity.com/paper/2110.02861/) (2021)
- [MInference 1.0: Accelerating Pre-filling for Long-Context LLMs via Dynamic Sparse Attention](https://mlsanity.com/paper/2407.02490/) (2024)
- [∞Bench: Extending Long Context Evaluation Beyond 100K Tokens](https://mlsanity.com/paper/2402.13718/) (2024)
- [Efficient Attentions for Long Document Summarization](https://mlsanity.com/paper/2104.02112/) (2021)
- [DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming Heads](https://mlsanity.com/paper/2410.10819/) (2024)
- [L-Eval: Instituting Standardized Evaluation for Long Context Language Models](https://mlsanity.com/paper/2307.11088/) (2023)
- [PyramidInfer: Pyramid KV Cache Compression for High-throughput LLM Inference](https://mlsanity.com/paper/2405.12532/) (2024)
- [Unlimiformer: Long-Range Transformers with Unlimited Length Input](https://mlsanity.com/paper/2305.01625/) (2023)
- [Keyformer: KV Cache Reduction through Key Tokens Selection for Efficient Generative Inference](https://mlsanity.com/paper/2403.09054/) (2024)
- [SparQ Attention: Bandwidth-Efficient LLM Inference](https://mlsanity.com/paper/2312.04985/) (2023)
- [One Thousand and One Pairs: A "novel" challenge for long-context language models](https://mlsanity.com/paper/2406.16264/) (2024)
- [NACL: A General and Effective KV Cache Eviction Framework for LLMs at Inference Time](https://mlsanity.com/paper/2408.03675/) (2024)

## Built on by (most cited first, 11 of 11)

- [Text or Pixels? It Takes Half: On the Token Efficiency of Visual Text Inputs in Multimodal LLMs](https://mlsanity.com/paper/2510.18279/) (2025)
- [LongFlow: Efficient KV Cache Compression for Reasoning Models](https://mlsanity.com/paper/2603.11504/) (2026)
- [LouisKV: Efficient KV Cache Retrieval for Long Input-Output Sequences](https://mlsanity.com/paper/2510.11292/) (2025)
- [Towards Efficient Large Language Model Serving: A Survey on System-Aware KV Cache Optimization](https://mlsanity.com/paper/2607.08057/) (2026)
- [PromptDistill: Query-based Selective Token Retention in Intermediate Layers for Efficient Large Language Model Inference](https://mlsanity.com/paper/2503.23274/) (2025)
- [Slow-Fast Inference: Training-Free Inference Acceleration via Within-Sentence Support Stability](https://mlsanity.com/paper/2603.12038/) (2026)
- [SeDeM: Selective Decompression of Hidden-State Memories for Long-Context Question Answering](https://mlsanity.com/paper/2608.00311/) (2026)
- [Random Attention: Rethinking KV Cache Eviction for Efficient Reasoning](https://mlsanity.com/paper/2609.03430/) (2026)
- [Attention Routing Stabilizes Early: Working-Set Inference for Recurrent Language Models](https://mlsanity.com/paper/2609.27373/) (2026)
- [MentorPulse: Refreshing Cross-Model Latent Guidance for Long-Form Generation](https://mlsanity.com/paper/2608.20927/) (2026)
- [On-Demand Attention: Language Models Know When to Recall](https://mlsanity.com/paper/2609.20734/) (2026)

## Abstract (arXiv)

Generating long sequences of tokens given a long-context input is a very compute-intensive inference scenario for large language models (LLMs). One prominent inference speed-up approach is to construct a smaller key-value (KV) cache, relieving LLMs from computing attention over a long sequence of tokens. While such methods work well to generate short sequences, their performance degrades rapidly for long-form generation. Most KV compression happens once, prematurely removing tokens that can be useful later in the generation. We propose a new inference method, RefreshKV, that flexibly alternates between full context attention and attention over a subset of input tokens during generation. After each full attention step, we update the smaller KV cache based on the attention pattern over the entire input. Applying our method to off-the-shelf LLMs achieves comparable speedup to eviction-based methods while improving performance for various long-form generation tasks. Lastly, we show that continued pretraining with our inference setting brings further gains in performance.