# SepLLM: Accelerate Large Language Models by Compressing One Segment into One Separator

arXiv: https://arxiv.org/abs/2412.12094 · HTML: https://mlsanity.com/paper/2412.12094/

## Key facts

- Cited by 41 papers in the map (data through 14 Oct 2026).
- Builds on 35 of its 46 references that are in the map.
- First posted on arXiv 16 Dec 2024.

## Builds on (most cited first, 30 of 35)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Longformer: The Long-Document Transformer](https://mlsanity.com/paper/2004.05150/) (2020)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [PIQA: Reasoning about Physical Commonsense in Natural Language](https://mlsanity.com/paper/1911.11641/) (2019)
- [The Pile: An 800GB Dataset of Diverse Text for Language Modeling](https://mlsanity.com/paper/2101.00027/) (2020)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention](https://mlsanity.com/paper/2006.16236/) (2020)
- [Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling](https://mlsanity.com/paper/2304.01373/) (2023)
- [Big Bird: Transformers for Longer Sequences](https://mlsanity.com/paper/2007.14062/) (2020)
- [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://mlsanity.com/paper/1912.08777/) (2019)
- [The LAMBADA dataset: Word prediction requiring a broad discourse context](https://mlsanity.com/paper/1606.06031/) (2016)
- [SnapKV: LLM Knows What You are Looking for Before Generation](https://mlsanity.com/paper/2404.14469/) (2024)
- [Crowdsourcing Multiple Choice Science Questions](https://mlsanity.com/paper/1707.06209/) (2017)
- [Compressive Transformers for Long-Range Sequence Modelling](https://mlsanity.com/paper/1911.05507/) (2019)
- [The Falcon Series of Open Language Models](https://mlsanity.com/paper/2311.16867/) (2023)
- [Linear Transformers Are Secretly Fast Weight Programmers](https://mlsanity.com/paper/2102.11174/) (2021)
- [Model Tells You What to Discard: Adaptive KV Cache Compression for LLMs](https://mlsanity.com/paper/2310.01801/) (2023)
- [PyramidKV: Dynamic KV Cache Compression based on Pyramidal Information Funneling](https://mlsanity.com/paper/2406.02069/) (2024)
- [Retrieve, Program, Repeat: Complex Knowledge Base Question Answering via Alternate Meta-learning](https://mlsanity.com/paper/2010.15875/) (2020)
- [InfLLM: Training-Free Long-Context Extrapolation for LLMs with an Efficient Context Memory](https://mlsanity.com/paper/2402.04617/) (2024)
- [PyramidInfer: Pyramid KV Cache Compression for High-throughput LLM Inference](https://mlsanity.com/paper/2405.12532/) (2024)

## Built on by (most cited first, 30 of 41)

- [Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention](https://mlsanity.com/paper/2502.11089/) (2025)
- [A Survey on Large Language Model Acceleration based on KV Cache Management](https://mlsanity.com/paper/2412.19442/) (2024)
- [MemGen: Weaving Generative Latent Memory for Self-Evolving Agents](https://mlsanity.com/paper/2509.24704/) (2025)
- [LightThinker: Thinking Step-by-Step Compression](https://mlsanity.com/paper/2502.15589/) (2025)
- [InftyThink: Breaking the Length Limits of Long-Context Reasoning in Large Language Models](https://mlsanity.com/paper/2503.06692/) (2025)
- [InfLLM-V2: Dense-Sparse Switchable Attention for Seamless Short-to-Long Adaptation](https://mlsanity.com/paper/2509.24663/) (2025)
- [SeerAttention-R: Sparse Attention Adaptation for Long Reasoning](https://mlsanity.com/paper/2506.08889/) (2025)
- [Mixture of Sparse Attention: Content-Based Learnable Sparse Attention via Expert-Choice Routing](https://mlsanity.com/paper/2505.00315/) (2025)
- [Thinking Inside the Mask: In-Place Prompting in Diffusion LLMs](https://mlsanity.com/paper/2508.10736/) (2025)
- [PackCache: A Training-Free Acceleration Method for Unified Autoregressive Video Generation via Compact KV-Cache](https://mlsanity.com/paper/2601.04359/) (2026)
- [IVC-Prune: Revealing the Implicit Visual Coordinates in LVLMs for Vision Token Pruning](https://mlsanity.com/paper/2602.03060/) (2026)
- [SentenceKV: Efficient LLM Inference via Sentence-Level Semantic KV Caching](https://mlsanity.com/paper/2504.00970/) (2025)
- [Dual Latent Memory for Visual Multi-agent System](https://mlsanity.com/paper/2602.00471/) (2026)
- [GEM: Empowering LLM for both Embedding Generation and Language Understanding](https://mlsanity.com/paper/2506.04344/) (2025)
- [OrthoRank: Token Selection via Sink Token Orthogonality for Efficient LLM inference](https://mlsanity.com/paper/2507.03865/) (2025)
- [A3: Attention-Aware Accurate KV Cache Fusion for Fast Large Language Model Serving](https://mlsanity.com/paper/2511.17560/) (2025)
- [ReMatch: Boosting Representation through Matching for Multimodal Retrieval](https://mlsanity.com/paper/2511.19278/) (2025)
- [FASA: Frequency-aware Sparse Attention](https://mlsanity.com/paper/2602.03152/) (2026)
- [Homogeneous Keys, Heterogeneous Values: Exploiting Local KV Cache Asymmetry for Long-Context LLMs](https://mlsanity.com/paper/2506.05410/) (2025)
- [G-KV: Decoding-Time KV Cache Eviction with Global Attention](https://mlsanity.com/paper/2512.00504/) (2025)
- [Attention Sink Forges Native MoE in Attention Layers: Sink-Aware Training to Address Head Collapse](https://mlsanity.com/paper/2602.01203/) (2026)
- [SemantiCache: Efficient KV Cache Compression via Semantic Chunking and Clustered Merging](https://mlsanity.com/paper/2603.14303/) (2026)
- [Memory Retrieval and Consolidation in Large Language Models through Function Tokens](https://mlsanity.com/paper/2510.08203/) (2025)
- [LightThinker++: From Reasoning Compression to Memory Management](https://mlsanity.com/paper/2604.03679/) (2026)
- [What to Keep, What to Forget: A Rate--Distortion View of Memory Compaction in LLMs and Agents](https://mlsanity.com/paper/2607.08032/) (2026)
- [Towards Efficient Large Language Model Serving: A Survey on System-Aware KV Cache Optimization](https://mlsanity.com/paper/2607.08057/) (2026)
- [Self-Adjust Softmax](https://mlsanity.com/paper/2502.18277/) (2025)
- [Sentence-Anchored Gist Compression for Long-Context LLMs](https://mlsanity.com/paper/2511.08128/) (2025)
- [SepSeq: A Training-Free Framework for Long Numerical Sequence Processing in LLMs](https://mlsanity.com/paper/2604.07737/) (2026)
- [Neural Attention Search](https://mlsanity.com/paper/2502.13251/) (2025)

## Abstract (arXiv)

Large Language Models (LLMs) have exhibited exceptional performance across a spectrum of natural language processing tasks. However, their substantial sizes pose considerable challenges, particularly in computational demands and inference speed, due to their quadratic complexity. In this work, we have identified a key pattern: certain seemingly meaningless separator tokens (i.e., punctuations) contribute disproportionately to attention scores compared to semantically meaningful tokens. This observation suggests that information of the segments between these separator tokens can be effectively condensed into the separator tokens themselves without significant information loss. Guided by this insight, we introduce SepLLM, a plug-and-play framework that accelerates inference by compressing these segments and eliminating redundant tokens. Additionally, we implement efficient kernels for training acceleration. Experimental results across training-free, training-from-scratch, and post-training settings demonstrate SepLLM's effectiveness. Notably, using the Llama-3-8B backbone, SepLLM achieves over 50% reduction in KV cache on the GSM8K-CoT benchmark while maintaining comparable performance. Furthermore, in streaming settings, SepLLM effectively processes sequences of up to 4 million tokens or more while maintaining consistent language modeling capabilities.